evolutionary histories of the marine fauna of southern Australia and yet this provides .... historical, geological and ocean circulation events in the Pacific and Atlantic basins. ...... AC -3â² and Fish R1 5â²- TAG ACT TCT GGG TGG CCA AAG AAT CA -3â²). ..... Each allele class (bin) was named by the modal integer and bins.
ASPECTS OF THE EVOLUTIONARY HISTORY OF A PAIR OF FISH SPECIES (ARRIPIDAE: ARRIPIS) ON EITHER SIDE OF A BIOGEOGRAPHIC BARRIER IN SOUTHERN AUSTRALIAN SEAS
Glenn Moore BSc. (Hons.) MSc.
This thesis is presented for the degree of
Doctor of Philosophy at Murdoch University 2012
DECLARATION
I declare that all sources are acknowledged and that this thesis is my own account of my research and contains, as its main content, work that has not previously been submitted for a degree at any tertiary education institution. The thesis contains no material previously written by another person except where due reference is made in the thesis itself.
___________________ Glenn Moore
For Dad, Mum & Debra
Nothing in biology makes sense except in the light of evolution. Theodosius Dobzhansky
I am turned into a sort of machine for observing facts and grinding out conclusions. Charles Darwin
Abstract Pairs of closely related species on either side of a barrier provide an opportunity to test evolutionary hypotheses. In southern Australian seas, many species pairs occur on either side of the Bass Strait, which temporarily becomes an isthmus during glacial stages, yet few evolutionary questions have been explored using these species. This thesis used partial nucleotide sequences from mitochondrial cytochrome b and COI genes and length-polymorphism in several nuclear introns in one such species pair, Arripis trutta (east of Bass Strait) and A. truttaceus (west of Bass Strait), to address some evolutionary questions. Data from the two remaining members of the family Arripidae, A. xylabion and A. georgianus, were included to provide context for interpreting patterns in the focal species pair. This research found no evidence for population subdivision in each of A. trutta, A. truttaceus and A. georgianus across their Australian distributions. This likely reflects panmixia, maintained by highly migratory life cycles and restricted breeding distributions. The molecular data also supported the currently recognised four-species taxonomy of the family. Phylogenetic reconstructions did not support the view of a sister-species relationship between the east-west species pair of A. trutta and A. truttaceus but instead suggested that these two species, together with A. xylabion, likely arose as a trifurcation (some 700,000 years ago), and so can be considered as a trio of sibling-species. All Arripis species were characterised by low genetic diversity and there was evidence for a more recent and/or severe population contraction in A. truttaceus compared to A. trutta. These different demographic signals support the a priori hypothesis that environmental conditions to the west of the Bass Strait were more severe than to the east during glacial stages and, in particular, the last glacial maximum. The results are considered in the context of the historical marine biogeography of temperate Australia. i
Contents ACKNOWLEDGEMENTS
v
1.0 GENERAL INTRODUCTION 1.1 FAMILY ARRIPIDAE 1.1.1 Arripis trutta Eastern Australian Salmon 1.1.1.1 General Biology and Ecology 1.1.1.2 Growth and Age 1.1.1.3 Life History 1.1.2 Arripis truttaceus Western Australian Salmon 1.1.2.1 General Biology and Ecology 1.1.2.2 Growth and Age 1.1.2.3 Life History 1.1.3 Arripis xylabion Giant Kahawai 1.1.3.1 General Biology and Ecology 1.1.4 Arripis georgianus Australian Herring 1.1.4.1 General Biology and Ecology 1.1.4.2 Growth and Age 1.1.4.3 Life History 1.2 PREVIOUS GENETIC STUDIES ON ARRIPIS 1.3 APPROACH 1.3.1 Genetic Markers 1.3.1.1 Mitochondrial DNA 1.3.1.2 Nuclear DNA 1.3.2 Analytical Methods 1.4 AIM AND RATIONALE OF THESIS 1.5 OUTLINE OF THESIS
1 7 9 9 11 11 14 14 15 16 18 18 19 19 20 20 22 24 25 25 26 28 29 30
2.0 GENERAL METHODS 2.1 SAMPLING DESIGN 2.2 SAMPLE COLLECTION, PROCESSING AND STORAGE 2.3 DNA EXTRACTION 2.4 MITOCHONDRIAL DNA 2.4.1 PCR Amplifications 2.4.2 Sequencing 2.4.3 Trials involving other mtDNA markers 2.5 NUCLEAR DNA 2.5.1 PCR Amplifications 2.5.2 Fragment Analysis 2.5.3 Primer Selection and Optimisation 2.5.4 Allele Scoring 2.6 NUCLEAR DNA SEQUENCES
32 32 36 36 37 38 39 40 41 41 42 45 46 52
3.0 POPULATION GENETIC STRUCTURE 3.1 INTRODUCTION 3.1.1 Aims and Hypothesis 3.2 MATERIALS AND METHODS 3.2.1 Sampling 3.2.2 Genetic Assays 3.2.3 Statistical Analyses 3.2.3.1 Levels of Polymorphism 3.2.3.2 HWE, Linkage Disequilibrium and Cohort Effects 3.2.3.3 Population Differentiation 3.2.3.4 Pairwise Differentiation 3.2.3.5 Overall Differentiation 3.2.3.6 Factorial Correspondence Analysis 3.2.3.7 Power Analysis 3.2.3.8 Temporal Stability
54 54 56 57 57 57 58 58 58 61 61 62 62 63 64
ii
3.3 RESULTS 3.3.1 Polymorphism, HWE and Linkage Disequilibrium 3.3.2 Cohort Effects 3.3.3 Population Differentiation 3.3.4 Power Analysis 3.3.5 Temporal Stability 3.4 DISCUSSION 3.4.1 Conclusions
64 64 72 72 73 73 81 85
4.0 A MOLECULAR APPROACH TO THE TAXONOMY OF AUSTRALIAN SALMON 4.1 INTRODUCTION 4.1.1 Aim 4.2 MATERIALS AND METHODS 4.2.1 Sampling Regime 4.2.2 Genetic Assays 4.2.3 Data Analyses 4.2.3.1 mtDNA Sequence Characteristics 4.2.3.2 Monophyly of mtDNA Sequences using Phylogenetic Analyses 4.2.3.2.1 Choice of outgroup 4.2.3.2.2 Concatenated sequence data 4.2.3.2.3 Substitution saturation 4.2.3.2.4 Substitution model 4.2.3.2.5 Phylogenetic analyses 4.2.3.2.5.1 Distance-based method 4.2.3.2.5.2 Maximum parsimony 4.2.3.2.5.3 Maximum likelihood 4.2.3.2.5.4 Bayesian inference 4.2.3.2.6 Parsimony networks 4.2.3.3 Fixed Differences Between mtDNA Sequences 4.2.3.4 Genetic distances based on mtDNA sequences 4.2.3.5 Fixed Differences in nDNA Genotypes 4.2.3.6 Differentiation in nDNA Allele Frequencies 4.2.3.7 Factorial Correspondence Analysis using Multilocus nDNA Genotypes 4.3 RESULTS 4.3.1 Mitochondrial DNA 4.3.1.1 Monophyly of mtDNA Sequences using Phylogenetic Analyses 4.3.1.2 Fixed Differences Between mtDNA Sequences 4.3.1.3 Intraspecific versus Interspecific Genetic Distances in mtDNA 4.3.2 Nuclear DNA 4.4 DISCUSSION 4.4.1 Conclusions
86 86 90 91 91 93 93 93 94 94 94 95 96 97 97 98 99 100 102 103 103 103 103 104 104 104 108 112 112 113 115 118
5.0 MOLECULAR PHYLOGENETICS, SPECIATION AND BIOGEOGRAPHY 5.1 INTRODUCTION 5.1.1 Objective 5.2 MATERIALS AND METHODS 5.2.1 Data 5.2.2 Data Analyses 5.2.2.1 Relationships Among Species of Australian Salmon 5.2.2.1.1 Testing between alternative tree topologies 5.2.2.1.2 Intron length polymorphism 5.2.2.2 Exploring the Phylogenetic Signal 5.2.2.2.1 Testing for a polytomy 5.2.2.2.2 Likelihood mapping 5.2.2.2.3 NeighborNet (Split Networks) 5.2.2.3 Molecular Dating 5.2.2.3.1 Testing for a clock-like rate of substitution 5.2.2.3.2 Divergence rate 5.2.2.3.3 Time to most recent common ancestor (Tmrca) 5.3 RESULTS 5.3.1 Relationships Among Species of Australian Salmon
119 119 120 121 121 122 122 122 123 124 125 125 126 127 127 128 128 129 130
iii
5.3.1.1 Support for Alternative Tree Topologies 5.3.2 Exploring the Phylogenetic Signal 5.3.3 Molecular Dating 5.4 DISCUSSION 5.4.1 Hard Polytomy in Australian Salmon 5.4.2 Pleistocene Divergence 5.4.3 Speciation in Australian Salmon 5.4.4 Conclusions
135 136 139 141 141 144 145 147
6.0 GENETIC DIVERSITY AND RECENT EVOLUTIONARY HISTORY 6.1 INTRODUCTION 6.1.1 Pleistocene Glacial/Interglacial Cycles in Southern Australia 6.1.1.1 East of the Bass Strait 6.1.1.2 West of the Bass Strait 6.1.2 Using Genetic Data to Infer the Evolutionary Histories of Arripis Species 6.1.3 Aim 6.2 MATERIALS AND METHODS 6.2.1 Sampling Regime 6.2.2 Genetic Assays 6.2.3 Genealogical Relationships 6.2.4 Data Analyses 6.2.4.1 Genetic Diversity 6.2.4.2 Comparisons of Genetic Diversity Between Species 6.2.4.3 Neutrality of Markers 6.2.4.4 Demographic History 6.2.4.4.1 Tests for evidence of changes in population size 6.2.4.4.1.1 Mitochondrial DNA sequences 6.2.4.4.1.2 Nuclear DNA allele frequencies 6.2.4.4.2 Estimates of timing of population expansions 6.2.4.4.3 Estimates of effective population sizes 6.3 RESULTS 6.3.1 General Information 6.3.1.1 Mitochondrial DNA Sequences 6.3.1.2 Nuclear DNA Allele Frequencies 6.3.1.3 Genealogical Relationships 6.3.2 Genetic Diversity 6.3.2.1 General 6.3.2.2 Comparisons of Genetic Diversity Between Species 6.3.3 Neutrality of Markers 6.3.4 Demographic History 6.3.4.1 Tests for Evidence of Changes in Population Size 6.3.4.1.1 Mitochondrial DNA sequences 6.3.4.1.2 Nuclear DNA allele frequencies 6.3.4.2 Estimates of Timing of Population Expansions 6.4 DISCUSSION 6.4.1 Genetic Diversity in Arripis 6.4.2 Recent Evolutionary History of Arripis trutta and Arripis truttaceus 6.4.3 Comparison of Arripis truttaceus and Arripis georgianus 6.4.4 Recent Evolutionary History of Arripis xylabion 6.4.5 Conclusions
148 148 149 149 152 153 153 154 154 155 155 156 156 158 158 159 159 159 161 162 163 163 163 163 164 166 168 168 170 171 173 173 173 175 178 179 179 182 185 186 187
7.0
188
GENERAL CONCLUSIONS
REFERENCES
192
APPENDICES A.1 mtDNA primers used during primer development A.2 EPIC-PCR primers tested in this study A.3 Technical issues for the use of fragment length polymorphism in EPIC markers
219 224 227
iv
Acknowledgements Firstly, I thank my supervisor, Dr. Jennie Chaplin. Jennie’s support, advice, guidance, encouragement and enthusiasm have been exceptional.
Perhaps most importantly,
Jennie has taught me to think, to learn and to persevere. Throughout, Professor Ian Potter has provided encouragement and financial support for my project. The Centre for Fish and Fisheries Research and all its members also gave me support and advice. I enjoyed sharing the lab with my friends Nicole Philips, Ertuğ Sezmiş, Michelle Gardner and Bryn Farmer, and they all provided a great deal of support, advice, help, laughter and an understanding ear when things just weren’t going to plan. I especially thank Ertuğ for suggesting EPIC-PCR as a way to include nuclear data in my project. I am indebted to Masaki Miya, Natural History Museum, Chiba, Japan for providing the complete mitochondrial genome of Arripis trutta.
I thank Frances Brigg (State
Agricultural Biotechnology Centre) who ran most of my sequences and fragment analyses. Frances’ knowledge and problem solving skills rescued my project on many occasions. My friends Dr. Michael Craig, Drs. Roddy and Marie Hale kept me going with words of experience, encouragement and advice, and crucial moments of escape. This project relied on the assistance and generosity of many people, primarily in providing and/or organising samples: Dumpy Wheatcroft (Albany Seafoods, Albany); Alan & Peta Miles (Selim Fisheries, Dunsborough); Steeg Hoeksema, Ben Chuwen, Dave Fairclough, George Young (Centre for Fish and Fisheries Research); Kyle and Yuki Armstrong (South Australia), Roddy Hale (Lincoln University), Marie Hale (University of Canterbury), Ian Kerr and Sallyann Gudge (Lord Howe Island Marine Park), Kevin Grickage, Rod Grickage and many anonymous recreational fishers. Fellow Arripis researchers, John Stewart and Julian Hughes (NSW Department of Primary Industries) kindly sourced all of the A. trutta from eastern Australia and prepared the tissue before sending to me. Mum, Natalie, my extended family, and my dogs Basie and Duke have helped me through with constant love and support. Finally, my biggest support came from my wife Debra who endured the journey with me, pretended to understand my explanations of neutral theory, coalescence and Bayesian inference, and survived being a ‘Ph.D. widow’ with incredible patience and love. v
Chapter 1. General Introduction
1.0
General Introduction
The marine fauna of temperate Australia is characterised by very high endemism, with most species restricted to the Southern Australian bioregion (Wilson and Allen 1987; Ponder 1995; Edgar 2000). For example, in temperate inshore fishes, 85% of the ~ 600 species are endemic, with a further 11% restricted to Australian and New Zealand waters (Wilson and Allen 1987; Hutchins 1994). Endemism at the generic level is also remarkably high in several groups, including fishes, molluscs and echinoderms, and probably more for which data are currently limited (Wilson and Allen 1987). Despite awareness of these unique communities, very little is known about the underlying evolutionary histories of the marine fauna of southern Australia and yet this provides the key to the origin of faunal characteristics of the region.
One of the most intriguing faunal characteristics of southern Australian seas is the reported presence of a large number and diverse range of geminate species to the east and west of the Bass Strait. (N.B. Geminate species are defined as “... twin species [and specifically, each other’s closest relatives, i.e. sister-species] - each one representing the other on opposite sides of some form of barrier”; Jordan 1908, p. 75). For example, at least 30 pairs of sister-species of fish alone are thought to be represented on either side of the Bass Strait (Hutchins 1994) and other groups such as macro-algae (Womersley 1981) and a range of invertebrates (Whitley 1932; Bennett and Pope 1953; Dartnall 1974; O’Hara and Poore 2000) are similarly believed to contain large numbers of such species pairs.
However, this observation is derived largely from records of the
distributions of morpho-species and only rarely has the geminate status of the eastern and western species been confirmed with molecular evidence. Such confirmation is 1
Chapter 1. General Introduction important because, in some cases (pertaining to other regions), molecular data have shown that members of a presumed geminate are actually most closely related to other species on the same side of the barrier, rather than to each other (e.g. Jones et al. 2003; Craig et al. 2004; Marko and Moran 2009; Miura et al. 2010).
The presence of many geminate species to the east and west of the Bass Strait, if confirmed, provides strong circumstantial support for the view that the Bass Strait is one of the most significant marine biogeographical features of the temperate Australian region. Other evidence includes the occurrence of a range of species and community limits in the general vicinity of this feature (Whitley 1932; Bennett and Pope 1953; Dartnall 1974; Womersley 1981; Hutchins 1994; O’Hara and Poore 2000; Hidas et al. 2007). Speciation in southern Australia has often been linked to repeated glacial cycles that affected the area with substantial fluctuations in climate and sea level (Dartnall 1974; Hutchins 1987; Howard and Prell 1992; Chenoweth et al. 1998; Petit et al. 1999; Edgar 2000; Hewitt 2000; Lambeck et al. 2002). During glacial maxima, sea levels were up to 140 metres lower than present, resulting in the emergence of a land bridge, the Bassian Isthmus, that extended from mainland Australia to southern Tasmania (Wells and Okada 1996; Lambeck et al. 2002). The Bassian Isthmus is widely believed to have subdivided a broad range of taxa and ultimately, in some cases, resulted in geminate species (Figure 1.1). In addition, during glacial events a cold-water front (Subtropical Convergence Zone) moved northwards, causing other large-scale changes to sea temperatures and current flows (Wells and Okada 1996; see Chapter 6), which may also have been pivotal in speciation processes (e.g. Burridge 2000b). It is often assumed that the timing of these speciation events is linked to glaciation events in the
2
Chapter 1. General Introduction Pleistocene (e.g. Dartnall 1974; Hutchins 1987; Chenoweth et al. 1998; Edgar 2000), although this has rarely been tested (but see Burridge 2000b, for example).
Figure 1.1
Hypothetical
effect
of
glaciation on the distribution of temperate Australian marine fauna west (grey) and east (black) of the Bass Strait. A. The Bassian Isthmus land-bridge was exposed by lower sea levels during glaciation (black
outline,
shown
for
southern
Australia only).
The distribution of
temperate
was
fauna
also
shifted
northward by a cold water front. B. Post glacial range shifts resulted in present day distributions of pairs of species and populations divided at the Bass Strait. After Waters et al. (2005).
The evolutionary implications of marine biogeographical barriers are probably best known in relation to the Isthmus of Panama (e.g. Bermingham et al. 1997; Knowlton and Weigt 1998; Williams et al. 2001; Arbogast et al. 2002; Marko 2002; McCartney and Lessios 2002; Lessios 2008; Hurt et al. 2009; Miura et al. 2010). However, unlike this isthmus, which has permanently isolated communities on either side, many barriers in marine systems, e.g. the Bassian Isthmus, the Florida Peninsula, Baja California, the Sunda Shelf, are in the form of a temporary and/or incomplete peninsula, often subject 3
Chapter 1. General Introduction to repeated inundations and exposures during sea level changes.
Studies on the
evolutionary implications of these temporary or incomplete barriers are limited (e.g. based on limited taxonomic coverage; see Bowen and Avise 1990; Burridge 2000b; Bernardi et al. 2003; Rocha and Bowen 2008) but have raised some interesting considerations regarding the mode and tempo of marine speciation, as follows.
Firstly, even such temporary barriers are sufficient to cause speciation, at least in some circumstances (e.g. as evidenced by multiple sets of sibling species on either side of the barrier). Secondly, the fact that speciation can occur in the face of incomplete or temporary barriers adds credence to the view that parapatric (rather than allopatric) speciation may be very common in the marine environment (Gavrilets 2003; Rocha and Bowen 2008).
Thirdly, the distribution of some taxa around these incomplete
biogeographic barriers can be explained by dispersal events, often associated with postglacial range expansion (e.g. Bernardi and Lape 2005), providing an alternative to a strictly vicariant regional model enforced in studies of the complete barrier of the Isthmus of Panama. Fourthly, these barriers are making an ongoing contribution to speciation (e.g. as evidenced by the presence of taxa at various stages and ages of divergence; see Bowen and Avise 1990; Bernardi et al. 2003). Finally, contrary to expectations, most of the speciation events are estimated to have occurred in the Pliocene or Miocene rather than in the Pleistocene (Rocha and Bowen 2008, although see Rocha et al. 2005; Robertson et al. 2006). Nevertheless, as noted above, studies of the evolutionary implications of temporary/incomplete land-bridges are limited to a few species at a few sites, and so additional studies are required to investigate the generality of these results.
4
Chapter 1. General Introduction In the context of testing hypotheses about the Bass Strait/Bassian Isthmus, studies that have used molecular data to assess the relationships among putative geminate species specifically, are essentially limited to Burridge (2000b; based on data in Burridge 1999; Burridge and White 2000; Burridge 2000a) and Waters et al. (2004). The former study used mitochondrial DNA sequence data to examine the relationships among species pairs in three genera of cirrhitoid fishes. In two of the three genera, the results were consistent with vicariant speciation associated with changes in oceanographic conditions in southern Australia during glaciation, but at different times, with one pair estimated to have diverged in the Miocene (7.8 – 11.2 Mya) and the other in the Pliocene (2.7 – 3.8 Mya). The third species pair was considered to have diverged by vicariance around the north of Australia (Burridge and White 2000; Burridge 2000b).
In the second example, Waters et al. (2004) used mtDNA and nDNA data to explore relationships among lineages of Patiriella (now Meridiastra) seastars, hypothesising, inter alia, that vicariance associated with the Bassian Isthmus isolating barrier was a major mechanism of speciation in the group. The authors demonstrated the existence of a pair of east-west distributed sister-species that was generally consistent with that hypothesis; they were estimated to have diverged in the Pliocene around 2.1 – 2.4 Mya.
In addition, there is a handful of studies that have used molecular data to investigate the relationships between populations of a single species to the east and the west of the Bass Strait, either incidentally or as a deliberate attempt to examine the role of the Bassian Isthmus (and associated oceanographic changes) in generating allopatric divergence (e.g. Brown 1991; Nurthen et al. 1992; Colgan and Paxton 1997; Kassahn et al. 2003; Waters and Roy 2003; Dawson 2005; Waters et al. 2005; York et al. 2008; Ayre et al. 5
Chapter 1. General Introduction 2009; Fraser et al. 2009; Coleman et al. 2011). The results of these intraspecific studies are also consistent with the view that the Bass Strait/Bassian Isthmus is an important isolating barrier.
Once the relationship between apparently closely related species on either side of a barrier such as the Bassian Isthmus is established, these species also provide a ‘natural experiment’ to test hypotheses about the evolutionary implications of certain environmental conditions because they represent very similar genomes placed in different environments (Lessios 1998). For example, the evolution of different lifehistory strategies in allopatric sister-species may be driven by environmental factors such as localised differences in water conditions or habitats (e.g. Cambray 1994; Coleman et al. 2011). Moreover, these ‘natural experiments’ are particularly important in the context of elucidating the effects of glacial cycles on the demographic histories of coastal marine species. In such species, departures from mutation-drift equilibrium are typically explained in terms of glacial effects based on a posteriori explanations of the results (e.g. Avise 2000; Bowen et al. 2001; Beheregaray et al. 2002; Bernardi 2005; López et al. 2010).
To illustrate the importance of studying geminates, there are several investigations that have shown that the genetic signatures of demographic events may be shared between species, but may differ in strength and/or timing, which is related to differences in historical environments.
For example, geminate species of blenniid fishes in the
Atlantic and Mediterranean basins were both affected by temperature changes during glacial events, but the retention of diversity was greater in one of the pair due to the presence of refugia within its range (Domingues et al. 2008). In another example, 6
Chapter 1. General Introduction Lessios et al. (2001) found evidence for population expansions in trans-isthmian geminate sea urchins, but the date of the start of the expansion in each ocean may have differed by some 40,000 years. Similarly, Muss et al. (2001) accounted for contrasting population subdivision in trans-isthmian geminate fishes with reference to dissimilar historical, geological and ocean circulation events in the Pacific and Atlantic basins.
In terms of the Bassian Isthmus, there is considerable paleontological evidence that the effects of glaciation were more severe to the west of this biogeographical feature than to the east (see Chapter 6). Thus far, the only a priori test of this assertion used transisthmian populations of the kelp Durvillaea potatorum (Fraser et al. 2009). This work tended to support the idea that glacial conditions were worse on the west however, the data interpretation is predicated on the assumption that the populations to the west of the Bass Strait were extirpated during the last glacial maximum, and is complicated by the potential influence of contemporary gene flow and selection (Fraser et al. 2009). A closely related pair of species, rather than conspecific populations, would help to overcome some of these difficulties.
1.1
Family Arripidae
This study provides one of the first detailed investigations into the evolutionary history of a fish taxon in the Southern Australian bioregion. The family Arripidae was selected for this investigation for several reasons, as follows. Firstly, there is assumed geminate species with a representative on either side of the Bassian Isthmus, as well as two other congeners that permit results for the geminates to be considered in a broader context (Hutchins 1994). Secondly, the family is endemic to the region of interest (temperate 7
Chapter 1. General Introduction Australasian seas; Paulin 1993) and so presumably the entire history of the family has been shaped by factors specific to this region. Thirdly, the family has few species (four), which are large, well known and conspicuous (Paulin 1993). Thus, it is almost certain that complete taxonomic coverage of the family, at least in regard to the extant members, is included in this study. Finally, as detailed below, the biology of the family is relatively well known, providing the foundations for interpreting genetic patterns.
In order to interpret the recent evolutionary histories of the Arripis species, it is important to understand the biology of each species and accordingly, the following section contains a brief introduction to the family and the first detailed review of published and unpublished research into the biology of all members of the family.
The Arripidae is one of very few fish families that are endemic to temperate Australian and New Zealand seas (Wilson and Allen 1987). This thesis follows the taxonomy of Paulin (1993) and so recognises a single genus Arripis, with four species, and uses the nomenclature of Eschmeyer (2011) and Rees et al. (2006). Three of the species are large and morphologically similar and collectively referred to as ‘Australian Salmon’, while the remaining species is smaller and commonly referred to as ‘Australian Herring’. In Australian mainland waters, the family is represented by two species of Australian Salmon and the Australian Herring. The Eastern Australian Salmon (A. trutta) occurs on the east coast, while the Western Australian Salmon (A. truttaceus) and the Australian Herring (A. georgianus) occur on the south and west coasts (Figure 1.2). Arripis trutta also occurs in New Zealand waters, where it is known as Kahawai. The third Australian Salmon species, the Giant Kahawai (A. xylabion) occurs around several south-west Pacific islands and possibly also northern New Zealand (Figure 1.2). 8
Chapter 1. General Introduction
Figure 1.2
Approximate distributions of the four species of Arripidae: Arripis trutta
(yellow), A. truttaceus (pink), A. xylabion (orange), and A. georgianus (blue).
1.1.1
Arripis trutta (Forster 1801)
1.1.1.1
General Biology and Ecology
Eastern Australian Salmon
Arripis trutta occurs in temperate coastal waters of eastern Australia, from north of Sydney, New South Wales (occasionally into southern Queensland) to the eastern parts of Victoria and south to Tasmania (Stanley 1978; Paulin 1993; Hutchins and Swainston 1999; Figure 1.2). It is also found around the North Island and northern parts of the South Island of New Zealand (Paulin 1993). Adults form large schools that inhabit exposed beaches and rocky reefs (Stanley 1978), with some evidence that their movements may also include a more oceanic distribution (Malcolm 1960). Adults generally reach up to 89 cm in length and 7 kg in weight (Hutchins and Swainston 1999), although they may grow larger (Duffy and Petherick 1999). 9
Chapter 1. General Introduction Dietary studies indicate that adults of A. trutta are opportunistic, with around 65% of their diet being fish (especially clupeids) and the rest comprising crustaceans and some molluscs (Baker 1971; Webb 1973b; Stewart et al. 2011). Documented predators of the species are few but include bottlenose dolphins Tursiops truncatus (Gales et al. 1992), little penguins Eudyptula minor (Montague and Cullen 1987), great cormorants Phalacrocorax carbo and pied cormorants P. varius (Marchant and Higgins 1990). Anecdotal evidence from fishers indicates that other marine mammals and sharks also target them. Parasites, with potential deleterious effects, are known from A. trutta (Hutson et al. 2004; Bolton et al. 2006; Catalano and Hutson 2010; Catalano et al. 2010;2011).
A commercial fishery targeting A. trutta has existed for many decades in eastern Australia and New Zealand, landing some one million fish per year (Scandol et al. 2008; TPC 2009; Anon. 2010; Stewart et al. 2011) and the species has a wide recreational angling appeal, with more than a million individuals caught annually in Australia (Henry and Lyle 2003) and New Zealand (Anon. 2010). The economic importance of this species led to the commencement of monitoring research in the 1940s and considerable resources were committed to tag/recapture and growth studies throughout the species’ range after catches began to fall in the 1950s and 1960s (Fairbridge 1950; Malcolm 1959; Nicholls 1973; Stanley and Malcolm 1977; Stanley 1978;1988a). Arripis trutta is known to have been a reasonable component of historical and contemporary Australian and New Zealand indigenous fish catches (Leach et al. 1996; Anon. 2010; Waddell 2011).
10
Chapter 1. General Introduction 1.1.1.2
Growth and Age
Post-larval growth in A. trutta varies slightly across the species distribution (Nicholls 1973; Stewart et al. 2011) and age estimates from otolith annuli predict a maximum age of around 12 years in Australia (Stewart et al. 2011), and possibly longer in New Zealand (Egglestone 1975). Age at maturity is around 2-4 years (Stanley and Malcolm 1977; Stewart et al. 2011).
1.1.1.3
Life History
The life history of A. trutta is characterised by a broad adult distribution and a highly dispersive pelagic larval stage followed by a juvenile stage that migrates from nursery areas to the adult distribution (Figure 1.3). Between April and October, most adults of A. trutta are concentrated in waters off New South Wales with quiescent gonads (Stanley 1978; Stewart et al. 2011). Adult aggregations are highly variable in size and distribution (e.g. Scandol et al. 2008) and anecdotal evidence suggests a recent increase in overall size and range (e.g. Miller 2007; Williams 2007). Extensive tagging and
Figure 1.3
(over page)
Generalised map of the distribution and movements of
Arripis trutta (dashed arrows), Arripis truttaceus (black arrows), and Arripis georgianus (grey arrows) in Australian waters. ‘I’ indicates presumed larval dispersal, including some local retention; ‘II’ represents direction of movement of juvenile fish; ‘III’ represents movements of sub-adult fish; and ‘IV’ represents adult distribution and movements. 11
Chapter 1. General Introduction otolith microchemistry studies indicate considerable movement of adult fish with no notable evidence for site-fidelity or fish returning to nursery areas (Stanley 1988a;1988c; Stewart et al. 2011), however, some non-breeding adults might be resident in Tasmanian waters (Stanley 1978; Hughes et al. 2011).
Spawning is concentrated in the northern parts of the species’ range in summer and early autumn, when the southward flowing East Australian Current and other associated currents are strongest (Nicholls 1973; Stanley and Malcolm 1977; Stewart et al. 2011), enabling transport of pelagic larvae to southern nursery areas. Spawning in A. trutta is believed to be concentrated in a relatively small area between Lakes Entrance, Victoria (37o51′S 148o01′E) and Bermagui, New South Wales (36o27′S 150o04′E) with some evidence that spawning time varies geographically (Stanley and Malcolm 1977; Stewart et al. 2011). Spawning is also known from New Zealand waters (Webb 1973a).
Gonad maturation in both sexes commences in October and continues through to March, although peak gonad maturation is in January (Stanley and Malcolm 1977; Stewart et al. 2011).
Ovarian development and maturation is not recorded from
Tasmanian fish, or most females from Victoria (Stanley and Malcolm 1977; Stewart et al. 2011). The released eggs of A. trutta are pelagic and 0.8 - 1.0 mm in diameter with a single oil globule (Bruce et al. 1998). Spermatogenesis take place in larger males in New South Wales and some parts of Victoria (Stanley and Malcolm 1977; Stewart et al. 2011).
Fertilisation is external and larvae hatch some 40 hours after fertilisation
(Stanley and Malcolm 1977) and are pelagic (Neira et al. 1997). The pelagic larval duration is unknown, but is probably around four to six months (Nicholls 1973; Robertson 1982; Stewart et al. 2011). 13
Chapter 1. General Introduction Larvae of A. trutta are transported south by the East Australian Current and postsettlement and first-year individuals are most abundant in shallow (nursery) areas in Victorian and Tasmanian waters, where they may be broadly sympatric with some postsettlement A. truttaceus (Stanley and Malcolm 1977; Stanley 1978; Figure 1.3). Through the juvenile life stage, fish gradually move northward through Victorian waters and into the adult distribution in New South Wales (Stanley 1978). Tagging studies indicate that adult fish make multi-directional movements, including into northern New South Wales (Stanley 1978; Stewart et al. 2011).
1.1.2
Arripis truttaceus (Cuvier 1829)
1.1.2.1
General Biology and Ecology
Western Australian Salmon
Arripis truttaceus occurs from north of Perth on the west coast of Australia to the western parts of Victoria and Tasmania (Gomon et al. 1994; Hutchins and Swainston 1999; Figure 1.2). As with A. trutta, adults form large schools along exposed beaches and rocky reefs (Cappo et al. 2000), possibly with some oceanic movements (Malcolm 1960).
Juvenile A. truttaceus utilise shallow bays and estuaries as nursery areas
(Malcolm 1960; Lenanton 1982; McCormack 2000). Arripis truttaceus is larger than A. trutta, reaching at least 96 cm in length and weighing in excess of 9 kg (Hutchins and Swainston 1999).
Adult A. truttaceus are piscivorous, feeding mainly on pelagic clupeoid species (Hoedt and Dimmlich 1994). Sub-adult fish in Victorian waters eat mostly pilchard Sardinops 14
Chapter 1. General Introduction neopilchardus, anchovy Engraulis australis and sandy sprat Hyperlophus vittatus, but the diet specialises slightly with age and, as fish form larger schools comprising subadult fish, they generally follow aggregations of pilchard (Hoedt and Dimmlich 1994). As juveniles, A. truttaceus feed on small benthic fishes, crustaceans and even the seagrass Zostera (Thompson 1957; Robertson 1982), with feeding success enhanced by the formation of small schools (Foster et al. 2001). The predators of A. truttaceus are likely to be the same as those described for A. trutta. Parasites, with potentially deleterious effects, are known from A. truttaceus (Catalano and Hutson 2010; Catalano et al. 2010;2011).
Similar to A. trutta, a commercial fishery targeting A. truttaceus has existed for decades in southern Australia, landing some one million fish per year (Smith and Brown 2007a; DPI 2008; Fowler et al. 2008; TPC 2009). This species is also a prized recreational angling species, with just under a million fish caught annually (Henry and Lyle 2003). As with its eastern congener, falling catch rates dating back to the 1950s led to a plethora of fisheries related research (Fairbridge 1950; Malcolm 1959;1960; Nicholls 1969;1971;1973;
Stanley
1978;
Kirkwood
and
Walker
1984;
Stanley
1986;1988b;1988c). This has resulted in one of the longest research histories for an Australian marine fish (Cappo et al. 2000).
1.1.2.2
Growth and Age
Arripis truttaceus is thought to grow faster than A. trutta (Fairbridge 1950; Nicholls 1971). Nicholls (1973) found that fish remaining in South Australia do not reach the mean length of most other four year old fish until they are five years old and reported 15
Chapter 1. General Introduction that growth rates in A. truttaceus decrease with increasing distance from the south-west spawning grounds. Additionally, these fish do not show gonad maturation (see below). Age at maturity was estimated at around 4 years (Nicholls 1973), however there is some uncertainty associated with this estimate because it was based on scale annuli, which may not reflect annual growth increments as accurately as otoliths (see Egglestone 1975; Stanley 1980), and the interpretation of the data was often confounded by the fact that A. truttaceus and A. trutta were considered to be a single species at the time.
1.1.2.3
Life History
The life history of A. truttaceus is similar to that of A. trutta (Figure 1.3). Between July and January, adult A. truttaceus concentrate around the Hopetoun and Esperance (~33o49′S 121o52′E) area of Western Australia with quiescent gonads (Malcolm 1960). Adult aggregations are highly variable in size (e.g. Smith and Brown 2007a). Tagging studies indicate considerable movement of adult fish with no evidence for site-fidelity or fish returning to nursery areas (Stanley 1986;1988b;1988c). Influences such as the distribution of prey (Hoedt and Dimmlich 1994) and oceanic conditions (Lenanton et al. 1991) are thought responsible for these movements.
In contrast to A. trutta, spawning in A. truttaceus occurs around the south-western corner of Australia in late autumn and early winter when the eastward flowing Leeuwin Current is strongest (Caputi et al. 1996), and eastward-directed winds are dominant (Petrusevics and Bye 1995), enabling transport of pelagic larvae to the south-east.
16
Chapter 1. General Introduction Gonad maturation in A. truttaceus commences in February and continues through June, when mature fish are found between Bremer Bay (34o22′S 119o22′E) and Perth in Western Australia (Malcolm 1960).
Peak gonad maturation and post-spawning
degeneration occurs in April and May, at which time the individuals concentrate in the south-west, from Cape Leeuwin (33o22′S 115o08′E) to Busselton (33o44′S 115o17′E). Malcolm (1960) argued that these observations, along with the fact that this is the only region in which fish in immediate post-spawning condition are landed, is evidence that spawning is centred in the region from Cape Leeuwin to Busselton.
During the
spawning phase, few adult fish are encountered further east (e.g. Esperance).
Ovarian development and maturation is completely absent in South Australian individuals, and using indirect evidence, Malcolm (1960) extended this status to females from Victoria and Tasmania. Spermatogenesis takes place in larger males in South Australia in phase with individuals in Western Australian, however there is no evidence that milt is ever released and testicular resorption is thought to occur (Malcolm 1960). Although the gametes and larvae of A. truttaceus are not specifically known, it is likely that they are very similar to those of A. trutta (see above). Arripis truttaceus larvae are believed to be transported away from the south-west spawning area toward the southeast by water flow because larval distribution appears to be positively correlated with both mean sea level along the south coast (as an indicator of the Southern Oscillation Index and hence, the Leeuwin Current) and the duration of easterly-directed winds (Lenanton et al. 1991; Petrusevics and Bye 1995). Pelagic larval duration is unknown, but is probably around four to six months (Nicholls 1973; Petrusevics and Bye 1995).
17
Chapter 1. General Introduction Post-settlement and first-year A. truttaceus occur along the entire southern coast of Australia but are most abundant in Victorian and Tasmanian waters, where they are broadly sympatric with post-settlement A. trutta (Malcolm 1960; Stanley and Malcolm 1977; Figure 1.3). As juveniles and sub-adults, they move away from south-easterly nursery areas to form larger schools (Robertson 1982; Hoedt and Dimmlich 1994), which make large multi-directional movements across their sub-adult range following schools of baitfish over three to four years. Following this immature growth stage, fish undertake a unidirectional pre-spawning/maturation migration to the adult distribution range (Figure 1.3).
Arripis truttaceus may well have the greatest documented
movement from juvenile to adult of any marine fish (Gillanders et al. 2003).
1.1.3
Arripis xylabion (Paulin 1993)
1.1.3.1
General Biology and Ecology
Giant Kahawai
Arripis xylabion has only recently been recognised (Paulin 1993) and, as a consequence of this and its isolated distribution (see below), little of its biology is known. Arripis xylabion is known from Lord Howe Island, Norfolk Island and the Kermadec Islands, with unconfirmed reports from northern New Zealand waters (Paulin 1993; Figure 1.2). It is a coastal pelagic schooling species, whose habits likely resemble those of A. trutta (Ian Kerr, Lord Howe Island Marine Park, pers. comm.). Arripis xylabion reaches at least 85 cm in length (Paulin 1993) and weighs up to 7 kg (Ian Kerr, pers. comm.). Since the other members of the Arripidae are generally similar to each other in their biology, it is likely that A. xylabion will conform to most of the biological attributes of the family, although the movements and dispersal patterns are difficult to predict. 18
Chapter 1. General Introduction 1.1.4
Arripis georgianus (Valenciennes 1831)
1.1.4.1
General Biology and Ecology
Australian Herring
The distribution of A. georgianus is broadly sympatric with A. truttaceus in most of its range, although it extends further north along the west coast of Australia to around Shark Bay (26o00′S 113o35′E) (Hutchins 1994; Fairclough et al. 2000a; Figure 1.2). The accuracy of the identification of specimens reported between eastern Victoria and Sydney is disputed (Hutchins and Swainston 1999; Gomon et al. 1994).
Arripis
georgianus are near-shore pelagic fish, which utilise estuaries more than other members of the family (Hutchins and Swainston 1999; Potter and Hyndes 1999).
Arripis
georgianus is the smallest member of the family, reaching a maximum size of 41 cm and weighing up to 1 kg (Hutchins and Swainston 1999).
The diet of adult A. georgianus is more opportunistic than A. truttaceus because, in addition to feeding on fish such as blue sprat Spratelloides robustus and sandy sprat Hyperlophus vittatus, their diet also commonly includes bivalves, amphipods, shrimps and crabs, as well as insects washed from terrestrial sources (Tregonning et al. 1995; pers. obs.). Fish in the 0+ age group rely on amphipods and other crustaceans and bivalves associated with detached macrophytes in the surf zone (Lenanton 1982). As per A. trutta and A. truttaceus, parasites, with potentially deleterious effects, are known from A. georgianus (Catalano and Hutson 2010; Catalano et al. 2010;2011).
A commercial fishery targeting A. georgianus, which lands 3 – 7 million fish per year, has existed for many decades in southern Australia (Tregonning et al. 1995; Smith and 19
Chapter 1. General Introduction Brown 2007b).
This species is also an extremely popular angling species that is
considered good eating (Hutchins and Swainston 1999). In fact, A. georgianus is the most heavily recreationally fished species in Australia, with around 7 million individuals taken annually (Henry and Lyle 2003). Although fisheries managers have long recognised the importance of research into this species (e.g. Lenanton and Hall 1976; Lenanton 1978), it was not until after catch rates started falling in the early 1990s that most studies were initiated (e.g. Tregonning et al. 1995; Fairclough et al. 2000a;2000b; Ayvazian et al. 2004).
1.1.4.2
Growth and Age
Consistent with the pattern seen in A. truttaceus, A. georgianus grow more slowly at the eastern end of their distribution compared to the west (Fairclough et al. 2000b), however fish do not show gonad maturation until reaching the south-west (see below). On average, female A. georgianus have higher growth rates, maximum size, maximum age, and age at maturity compared to males (Lenanton 1978; Fairclough et al. 2000a;2000b). Otolith studies show that A. georgianus live up to nine years, and reach sexual maturity by the end of their second year (Fairclough et al. 2000b).
1.1.4.3
Life History
Tagging studies indicate considerable movement of adults of A. georgianus, with no evidence for site-fidelity or for fish returning eastward to nursery areas (see Tregonning et al. 1995; Ayvazian et al. 2004). Non-spawning adults of A. georgianus form schools
20
Chapter 1. General Introduction along the lower west and south coasts of Australia (Tregonning et al. 1995; Smith and Brown 2007b).
Like the sympatric A. truttaceus, spawning in A. georgianus coincides with the eastward flowing Leeuwin Current (Caputi et al. 1996) and other eastward-directed winds (Petrusevics and Bye 1995), enabling transport of pelagic larvae to south-eastern nursery areas (Lenanton et al. 1991; Fairclough et al. 2000b; Ayvazian et al. 2004). Gonadosomatic indices for A. georgianus in the south-west of Australia show that peak maturation in males occurs around April/May and in females about a month later (Lenanton 1978; Fairclough et al. 2000a). As the spawning season commences, ripe adults merge into large schools (Tregonning et al. 1995; Fairclough et al. 2000a) throughout much, but not all, of their distribution in south-western Australia (between Bremer Bay and the Abrolhos Islands (28o43′S 113o46′E; Fairclough et al. 2000a; Figure 1.3). There is also some evidence that spawning might take place further north of this area (Shark Bay; Fairclough et al. 2000a), but testicular and ovarian development and maturation is completely absent in A. georgianus east of Esperance (Lenanton 1978; Fairclough et al. 2000a;2000b), and spawning does not occur in estuarinedwelling individuals of A. georgianus (S. Hoeksema, Murdoch University, pers. comm.). The released eggs of A. georgianus are pelagic and 0.4 - 0.6 mm in diameter (Lenanton 1978). Larvae are similar to those of other members of the family (Neira et al. 1997), and the pelagic larval duration is unknown, but is probably around two to four months (Lenanton et al. 1991; Fairclough et al. 2000b; Ayvazian et al. 2004).
Juvenile A. georgianus occur along the entire southern coast of Australia but are most abundant in seagrass beds in South Australian and Victorian waters (Tregonning et al. 21
Chapter 1. General Introduction 1995). While there is some anecdotal evidence for greater local retention of larvae in this species than for others in the family (Tregonning et al. 1995; Fairclough et al. 2000b), it was not supported by genetic data (Ayvazian et al. 2004). Fairclough et al. (2000b) demonstrated a gradual movement of growing fish from nursery areas in the south-east of Australia toward the south-west and predicted that, via this gradual migration, fish enter the adult distribution in their second year of life.
1.2
Previous Genetic Studies on Arripis
Only two previous studies have investigated the population genetics of one or more species of Arripis and only one of these has been published. Since they are highly relevant to the present research, details of these studies are considered below.
The first study was the PhD research of MacDonald (1980), which focussed on an assessment of the population structure of A. truttaceus, but also provided preliminary information about the phylogenetic relationships among A. trutta, A. truttaceus and A. georgianus (A. xylabion was not described at that time). Reflecting the era in which it was conducted, MacDonald’s work was based on allozyme markers. Apart from an abstract for a conference (MacDonald 1983), this work was never published. There were three main findings from MacDonald’s (1980) work.
Firstly, the levels of allozyme polymorphism in each of the three species of Arripis were generally low (Table 1.1) relative to those in most other marine teleosts (Selander 1976; Ward et al. 1994). MacDonald (1980) argued that the low levels of allozyme diversity in A. truttaceus, in particular, were indicative of a population bottleneck. 22
Chapter 1. General Introduction Secondly, MacDonald (1980) found no evidence of population subdivision in A. truttaceus based on samples from nine sites on the south coast between Port Phillip Bay (Melbourne, Victoria) and around Albany (Western Australia), which is consistent with the highly mobile life-style of this species (see Section 1.1.2.3). However, the test had limited power, primarily because it was based on the analysis of the patterns of variation at a single allozyme locus (EST), but also because no populations west of Albany were included. MacDonald (1980) did not include population level sampling for A. trutta or A. georgianus.
Thirdly, in a simple phylogenetic reconstruction, A. trutta and A. truttaceus were determined to be sister-species and tentatively estimated to have diverged some 0.8 to 1.14 Mya, and were linked to A. georgianus from an earlier split, around 5.8 to 6.7 Mya (MacDonald 1980).
The second study was an allozyme-based assessment of the population structure of A. georgianus that was conducted by Ayvazian et al. (2004), as part of a multi-disciplinary stock assessment.
Reflecting the results of MacDonald (1980) for A. truttaceus,
Ayvazian et al. (2004) also found a very low level of allozyme polymorphism in A. georgianus and no evidence of population subdivision, based on a test with very limited power (Table 1.1). Specifically, although the test was based on data from 23 allozyme loci, in each case the frequency of the most common allele exceeded 0.95. There is clearly a need to find more polymorphic markers to improve the power of the analyses. Thus far, no studies have used molecular markers to investigate the amount and distribution of genetic variation in any species of Arripis. However, the nucleotide sequence of a portion of the COI gene of small number of individuals of each Arripis 23
Chapter 1. General Introduction species has been determined as part of a taxonomic barcoding exercise (Fish Barcode of Life Initiative; Ward et al. 2005).
Allozyme polymorphism among n individuals of three species of Arripis
Table 1.1
reported by MacDonald (1980) and in A. georgianus by Ayvazian et al. (2004; data indicated by asterisk).
The proportion of polymorphic loci was calculated as the
number of loci with ≥ 2 alleles. The number of polymorphic loci for which the frequency of the common allele was ≤ 0.95 is also provided. Nei’s (1987) average expected heterozygosity (HE) was recalculated from the original data.
n Polymorphic loci (number assayed)
Number loci ≤ 0.95 HE
1.3
A. trutta
A. truttaceus
A. georgianus
A. georgianus*
103
580
20
646
0.37 (27)
0.22 (27)
0.23 (26)
0.26 (23)
4
1
4
0
0.06
0.01
0.04
0.02
Approach
The present study builds upon the previous genetic studies of Arripis species (i.e. MacDonald 1980; Ayvazian et al. 2004). It represents a significant step forward from these studies for the following reasons: (i) it jointly considers all four extant species of Arripis, although it focuses on A. trutta and A. truttaceus due to their putative geminate status; (ii) it takes advantage of advances in molecular techniques to directly sample variation at the level of the DNA molecule and so provides a greater amount of information than was possible from the allozyme markers; and (iii) it takes advantage of significant advances in approaches to phylogenetic and phylogeographic research. 24
Chapter 1. General Introduction 1.3.1
Genetic Markers
The importance of using multiple independent genetic markers to address population and evolutionary questions is widely recognised. This is because different markers have different attributes (e.g. inheritance and mutation rates), and concordance (or not) between markers can provide insights into evolutionary history that cannot be inferred from a single gene (Avise 2009). The present research is based upon two types of genetic markers; nucleotide sequence in mitochondrial DNA and length variation in the alleles of nuclear introns. The reasons for the selection of these markers are explained below.
1.3.1.1
Mitochondrial DNA
Mitochondrial DNA (mtDNA) is a small, circular DNA molecule in the mitochondrion of cells, and has been the mainstay of evolutionary genetics for decades (Brown et al. 1979; Avise 2004). Although debate continues to assess the strengths and weaknesses of the marker (e.g. Zink and Barrowclough 2008; Galtier et al. 2009), mtDNA is generally considered to be an ideal marker for resolving evolutionary questions (such as population subdivision and phylogeography) mainly due to its absence of recombination and uniparental (maternal) inheritance (e.g. Burridge 2000b; Birky 2001; Ward et al. 2005; Percides and Coelho 2006; Mateus et al. 2011). The mtDNA molecule comprises a number of different regions, some of which have different substitution rates and therefore provide the potential to examine processes over different timescales (Ballard and Whitlock 2004). However, the different mtDNA regions are all part of the same molecule and therefore rigidly linked, violating any assumption of independence 25
Chapter 1. General Introduction (Ballard and Whitlock 2004). Another limitation associated with the use of mtDNA is that such markers allow only the reconstruction of maternal lineages (Avise et al. 1987).
1.3.1.2
Nuclear DNA
Many of the limitations associated with using mtDNA markers to address evolutionary questions can be lessened by also incorporating nuclear DNA (nDNA) markers into the approach (e.g. Ballard and Whitlock 2004; Zink and Barrowclough 2008). This is advantageous for a number of reasons, including that the nDNA markers reflect both male and female contributions to the genetic composition of the population because they have biparental inheritance.
Nuclear DNA markers also provide ample scope for
repeated independent testing of a hypothesis or outcome because the nuclear genome contains a large number of unlinked loci and all nuclear genes are inherited independently of the mitochondrial genome (Avise 2004).
Microsatellite loci are probably the most commonly used type of nDNA marker in investigations into the population genetics and recent evolutionary history of single species or of closely related species.
However, while many of the properties of
microsatellite markers make them well suited for this role (reviewed in Li et al. 2002; Ellegren 2004), such use is associated with several drawbacks, including the fact that these markers typically have a narrow taxonomic range and the development of customised markers is relatively expensive and/or labour intensive and requires specialist knowledge and equipment (see Jarne and Lagoda 1996). These problems can be a major issue for studies like the present, which deal with species of low commercial and/or conservation value in a competitive funding environment. Consequently, it 26
Chapter 1. General Introduction would be a considerable advantage to find markers that exhibit many of the advantageous qualities of microsatellite markers (e.g. co-dominant markers with relatively high amounts of fragment length variation that can be ascribed to a particular locus and easily and unambiguously scored; Li et al. 2002), but which can be amplified using universal primers (and thereby avoid the time and expense of the developmental work of microsatellites).
Length variations in exon-primed, intron-crossing (EPIC) markers offer one such possibility. This technique uses ‘universal’ primers to anneal to a part of the exon of a gene (which should be relatively conserved) and thereby allow for the amplification of the adjacent part of the adjacent intron (which is not translated and therefore potentially more variable) (Palumbi and Baker 1994). The technique was initially used to amplify alleles that were then sequenced (e.g. Lessa 1992; Hare 2001; Atarhouch et al. 2003; Lavoué et al. 2003; Burridge et al. 2006; Morrison et al. 2006; Cooke and Beheregaray 2007), however, more recently, it has been used to examine length variation in the amplified intron sequence (e.g. Berrebi et al. 2005; Atarhouch et al. 2007; Fauvelot et al. 2007; Rolland et al. 2007). However, the number of studies that have used fragment length polymorphism in EPIC markers to address population genetic and/or evolutionary questions is very limited (but see above), and the full potential of this type of marker in these applications is not yet known. Some of the potential limitations associated with the use of EPIC markers in general and length polymorphism in these markers in particular have already been identified, although most of these are not specific to this type of marker (e.g. paralogy; Creer 2007). One of the main issues is the possibility of non-neutrality, given that these markers are a part of a gene and tightly linked to exons, although the empirical data typically do not bear this out (see Friesen 27
Chapter 1. General Introduction 2000; Creer 2007). Another issue is that the underlying mutational model for these markers is unknown (Friesen 2000; Lynch 2002; Creer 2007). In this regard, it is worth noting that even in the case of microsatellite markers, which are widely used in population and evolutionary studies and the properties of which are relatively well known, the mutational model is not fully understood and likely varies among species and loci (Estoup et al. 2002). Problems stemming from uncertainty about the mutational model of a class of marker can often be circumvented by repeating analyses using contrasting models (e.g. see Cornuet and Luikart 1996).
1.3.2
Analytical Methods
The primary goal of this research was to use genetic data to elucidate aspects of the origins and evolutionary histories of Arripis species (see below). The main questions being addressed fall within the fields of population genetics, phylogenetics and phylogeography. Each of these fields has undergone significant advances in recent times due to a combination of increased statistical rigor, associated with a capacity to generate relatively large amounts of genetic data and increased computing power, and significant conceptual gains (e.g. Excoffier et al. 2005; Beheregaray 2008; Avise 2009; Edwards 2009; Knowles 2009). The main advance in each of these fields is probably the ability to develop customised hypotheses (model-based perspective) and test these hypotheses with a relatively high degree of precision (see Knowles 2009; Beaumont et al. 2010). Having said that, there are still limitations in the application of genetic data to ecological and evolutionary questions. For example, the presence of inherently low polymorphism across a species genome still limits the strength of many evolutionary inferences (e.g. Ryman et al. 2006; Palm et al. 2009), although this is less problematic 28
Chapter 1. General Introduction than it once was (e.g. Brito and Edwards 2008; Edwards 2009; Knowles 2009; Heled and Drummond 2010). Furthermore, phylogenetic reconstructions for so-called ‘rapid radiations’ remain challenging, although this is being increasingly addressed (e.g. Huelsenbeck et al. 2002; Huson and Bryant 2006; Moulton and Huber 2009; Willerslev et al. 2009; Bloomquist et al. 2010; Heled and Drummond 2010).
1.4
Aim and Rationale of Thesis
This thesis examines the distribution of genetic variation within and between species in the family Arripidae and uses the resultant information to test the roles of paleogeographic and paleoclimatic conditions in shaping the evolutionary histories of these species. It focuses on two species, Arripis trutta (Eastern Australian Salmon) and A. truttaceus (Western Australian Salmon) in particular, as these represent putative geminate species. It also includes information on A. georgianus (Australian Herring), which has a similar (west coast) distribution to A. truttaceus, and on A. xylabion (Giant Kahawai) from Lord Howe Island, which on the basis of morphological evidence (see species account) appears to be very closely related to the two species of Australian Salmon. Information on each of the latter two species, although significant in its own right, is largely used to help place the results for A. trutta and A. truttaceus into context.
This research is significant for the following reasons.
Firstly, it extends existing
information about the population structure, taxonomy, relationships, origins and evolutionary history of members of the family Arripidae. Importantly, it is the first genetic study to include all four species, and both mtDNA and nDNA markers. Secondly, it provides one of the first investigations into the relationships (and status) of 29
Chapter 1. General Introduction putative geminates on the east and west coasts of Australia. Thirdly, it provides one of the first tests of the prediction that species which reside to the west of the Bass Strait may have been more severely impacted by glacial cycles than those that reside to the east. Therefore, this research represents an important early step in the development of knowledge about the origins and evolutionary histories of the biota of the Southern Australian bioregion.
1.5
Outline of Thesis
Chapter 1: General Introduction. The aim of the General Introduction (the present chapter) was to introduce the conceptual framework, provide general background information, state the significance and overall aim of the study and outline the aims of each chapter.
Chapter 2: General Material and Methods. The aim of Chapter 2 was to provide the general details of the methodology. It describes the sampling regime and methodology, as well as the genetic methods and includes an assessment of the reliability of the EPIC markers.
Chapter 3: Population Genetic Structure. The aim of Chapter 3 was to use genetic (intron length polymorphism) data to investigate the population genetic structures of A. trutta, A. truttaceus and A. georgianus effectively over their entire geographic distributions (latter two species) or the entire Australian mainland distributions (former species).
This was necessary because an understanding of the underlying genetic
30
Chapter 1. General Introduction structures of each of these species was essential for interpreting the results of the evolutionary aspects of this study.
Chapter 4: A Molecular Approach to the Taxonomy of Australian Salmon. The aim of Chapter 4 was to investigate whether the current, morphologically-based taxonomy of Australian Salmon is supported by DNA data. It was necessary because it is important to confirm that these taxa represent ‘true’ species, as opposed to conspecific populations that are exchanging genes or have recently done so.
Chapter 5: Molecular Phylogenetics, Speciation and Biogeography.
The aim of
Chapter 5 was to generate information on the relationships among the species of Australian Salmon and use the results to develop an explanation for the speciation and current biogeography of this group. It focussed on testing the assumption that the east/west species of A. trutta and A. truttaceus are true geminate species, and applied molecular dating to estimate divergence times between species of Australian Salmon.
Chapter 6: Genetic Diversity and Recent Evolutionary History. The aim of Chapter 6 was to assess and compare the levels and patterns of genetic diversity in the effectively geminate species A. trutta and A. truttaceus and use the resultant information to infer and compare aspects of the recent evolutionary histories of these species. The chapter tests the prediction that A. truttaceus, which resides to the west of the Bass Strait, has been more severely impacted by glacial cycles than A. trutta, which resides to the east.
Chapter 7: General Conclusions. The aim of Chapter 7 was to summarise the main findings of each chapter. 31
Chapter 2. General Methods
2.0
2.1
General Methods
Sampling Design
Sampling was designed to maximise the potential for detecting genetic heterogeneity both within and between Arripis species (Page and Holmes 1998). The design of the sampling regime considered the following. The consequences and limitations of each consideration in the context of each question are discussed in the relevant chapters.
Taxonomic coverage: all four extant species of Arripis were included in the thesis, although the extent of their inclusion varied due to the specific questions that were being asked (see Aim for each chapter), and access to samples (e.g. A. xylabion was not included in an assessment of population structure because all individuals were sampled from a single site; see Chapter 3).
Spatial coverage: Samples were obtained from virtually either the entire distribution of the species in Australian waters (A. trutta) or the entire distribution of the Australian species (A. truttaceus, A. georgianus). This strategy was used to reduce the possibility of missing genetic differentiation if, for example, all individuals had been sampled from a single, more localised breeding unit within a subdivided species (see Page and Holmes 1998). Due to a reliance on others for specimens, spatial sampling of A. xylabion was not possible (see below).
Temporal framework: In order to minimise the potential effects of temporal genetic variations (e.g. Johnson and Black 1984; Hedgecock 1994; Dannewitz et al. 2005; Palm 32
Chapter 2. General Methods et al. 2009), sampling for each species was generally completed within the same few months (Table 2.1). However, samples of A. trutta from New Zealand and samples of A. truttaceus and A. georgianus from South Australia proved difficult to source and were taken one to two years after the rest of the collections (Table 2.1). Extensive temporal sampling was not undertaken, partly because the initial analysis revealed no evidence of significant spatial genetic heterogeneity. However, a second sample of A. georgianus from Fremantle, Western Australia was collected two years after the first, to provide a very limited test for the presence of temporal genetic heterogeneity (see Section 3.2.1).
Sample size:
The sample sizes for each species from each sampling location are
indicated in Table 2.1. A minimum of 30 individuals was sought for each sampling location (e.g. Ryman et al. 2006; Pruett and Winker 2008). However, as described below, the project was largely reliant upon specimens collected by other individuals and this minimum number was not always obtained.
Samples of 29 – 30 A. trutta were collected from three locations throughout most of the geographic range of this species in Australian waters (Table 2.1, Figure 2.1).
In
addition, two specimens of this species were also obtained from near Christchurch, New Zealand.
Samples of 13 – 45 A. truttaceus were collected from five locations
throughout most of the geographic range of this species, which is confined to Australian waters (Table 2.1, Figure 2.1). Samples of 9 – 44 A. georgianus were collected from seven locations throughout most of the geographic range of this species, which is also confined to Australian waters (Table 2.1, Figure 2.1). A single sample of 25 A. xylabion was collected from Lord Howe Island (Table 2.1, Figure 2.1). 33
Table 2.1
Collection details for Arripis species including approximate coordinates of sampling locations. See also Figure 2.1.
ID (n) Arripis trutta EN (29) EV (30) ET (30) EZ (2) Arripis truttaceus WA (30) WY (15) WL (39) WW (13) WS (45) Arripis xylabion KL (25) Arripis georgianus HF (44) HF08 (41) HC (39) HI (23) HP (36) HO (28) HW (09) HS (41)
Site
Latitude
Longitude
Collection Date/s
Source
Port Stephens, NSW Lakes Entrance, Vic. Flinders Island, Tas. Christchurch, NZ
32° 40′ 31″ S 37° 52′ 48″ S 39° 59′ 49″ S 43° 31′ 48″ S
152° 12′ 27″ E 147° 59′ 24″ E 148° 02′ 28″ E 172° 37′ 13″ E
09 May 2007 31 May 2007 April-May 2007 30 January 2009
NSW DPI NSW DPI NSW DPI Commercial
Augusta (Hamelin Bay), WA Yeagarup Beach, WA Albany, WA Wellstead Estuary, WA The Coorong, SA
34° 12′ 25″ S 34o 26′ 28″ S 35° 00′ 35″ S 34° 23′ 17″ S 35° 56′ 41″ S
115° 02′ 08″ E 115o 37′ 39″ E 117° 55′ 31″ E 119° 23′ 05″ E 139° 25′ 37″ E
13 March 2007 29-31 March 2007 27 March 2007 March 2007 05 July 2008
Commercial Recreational Commercial Murdoch Commercial
Lord Howe Island, NSW
31° 33′ 10″ S
159° 04′ 50″ E
Aug. – Nov. 2008
Recreational
Fremantle, WA Fremantle, WA Canal Rocks, WA Iwin Inlet, WA Peaceful Bay, WA Albany (Oyster Harbour), WA Wellstead Estuary, WA The Coorong, SA
32° 02′ 43″ S 32° 02′ 43″ S 33° 40′ 10″ S 34° 59′ 30″ S 35° 02′ 30″ S 35° 03′ 12″ S 34° 23′ 17″ S 35° 56′ 41″ S
115° 44′ 07″ E 115° 44′ 07″ E 114° 59′ 40″ E 116° 57′ 30″ E 116° 55′ 52″ E 117° 53′ 28″ E 119° 23′ 05″ E 139° 25′ 37″ E
06-09 July 2006 24-26 April 2008 13-14 July 2006 May-August 2006 May-July 2006 May-August 2006 May-August 2006 05 July 2008
Recreational Recreational Recreational Murdoch Commercial Murdoch Murdoch Commercial
Figure 2.1
Approximate collection locations for samples of Arripis used in this study: A. trutta (yellow), A. truttaceus (pink), A.
xylabion (orange), A. georgianus (blue). See Table 2.1 for details.
Chapter 2. General Methods 2.2
Sample Collection, Processing and Storage
The Arripis samples were sourced from recreational and commercial fishers and from other research projects (at Murdoch University and at New South Wales Department of Primary Industries) between May 2006 and January 2009 (Table 2.1).
For the westerly distributed species (A. truttaceus and A. georgianus), whole or part fish were placed on ice or into liquid nitrogen as soon as possible after capture before being transported to the laboratory, where they were stored frozen until processed. Muscle tissue was sampled and stored at -20o C. Since they had to be transported by air, tissue samples from A. trutta and A. xylabion from New South Wales and Lord Howe Island, respectively, were shredded into very small fragments and placed into a 10% DMSO - 5 M NaCl solution (Proebstel et al. 1993; Kilpatrick 2002) at a ratio of approximately 1:50 tissue to solution and stored frozen. Upon arrival at Murdoch University, these tissue samples was transferred to a smaller volume (~ 1.5 ml) of the DMSO solution and stored at -20o C. DNA from each of two individuals of A. trutta from New Zealand was mailed to Murdoch University after being extracted by Marie Hale (University of Canterbury,
Christchurch)
using
Qiagen
extraction
columns
following
the
manufacturer’s protocol. They were mailed in 35 µl of TE Buffer (10 mM Tris-HCL [pH 8.0], 1 mM EDTA) and frozen at -20o C on arrival.
2.3
DNA Extraction
Total genomic DNA was isolated from individual tissue samples using a MasterPureTM extraction kit (Epicentre 2003). The A. trutta tissue was soaked in 200 µl of autoclaved 36
Chapter 2. General Methods water for approximately 1 minute, to remove much of the NaCl storage solution, and blotted dry before extraction. A 5 mg sample of tissue was homogenised in 300 µl of Tissue and Cell Lysis Solution and Proteinase K (50 µg) and incubated at 65o C for 15 minutes. Buffered RNase A (5 µg) was added to the homogenate, which was then incubated at 37o C for a further 30 minutes. While the homogenate was on ice, 175 µl of MPC Protein Precipitation Reagent was added and the mix was centrifuged for 10 minutes at 10,000 x g. The resultant supernatant was transferred to a new tube and DNA was precipitated by centrifuging for 10 minutes at 10,000 x g in 500 µl of isopropanol. The pellet was washed with two 500 µl aliquots of 75% ethanol. Once dry, the pellet was resuspended in 35 µl of TE Buffer (10 mM Tris-HCL [pH 8.0], 1 mM EDTA) and stored at -4o C. The presence (or otherwise) of DNA was confirmed by electrophoresis on a 2% agarose gel, stained with ethidium bromide or SYBR®Safe.
2.4
Mitochondrial DNA
Analysis of mtDNA variation in the four species of Arripis was based on the nucleotide sequence of a 483 bp portion (positions 556 - 1038 in Arripis) of the cytochrome b gene and a 654 bp portion of cytochrome oxidase subunit 1 (COI) (positions 52 – 705 in Arripis). These gene regions were selected for analysis because they could be amplified reliably from Arripis (unlike the control region; see Section 2.4.3) and are widely used in similar studies (e.g. Burridge and White 2000; Burridge 2000b; Ward et al. 2005; Percides and Coelho 2006). The cytochrome b partial gene sequences were amplified using Arripis specific primers, which were developed as a part of the present study using an alignment of unpublished sequence (Masaki Miya, Natural History Museum, Chiba, Japan, pers. comm.) and published sequence for A. trutta cytochrome b (Doiuchi 37
Chapter 2. General Methods and Nakabo 2006). The COI partial gene sequences were amplified using the universal primers of the fish barcoding project FISHBOL (Ward et al. 2005) that were already known to work in all species of Arripis (B. Ward, CSIRO, pers. comm.).
2.4.1
PCR Amplifications
Polymerase chain reaction (PCR) in a GeneAmp 9700 Thermal Cycler (Applied Biosystems) was used to amplify the target regions of the mtDNA, with conditions specific to each gene. The cycling conditions for the cytochrome b region were: (i) an initial denaturation phase of 5 minutes at 94˚ C; (ii) 30 amplification cycles, with each cycle consisting of 30 seconds of denaturation at 94° C, 30 seconds of annealing at 50° C, 30 seconds of extension at 72° C; and (iii) a final 5 minute extension at 72˚ C.
Each cytochrome b PCR reaction mixture contained 0.1 mM of each of the dNTPs (Promega), 0.25 U of Taq DNA polymerase (Roche), 5 µl Taq Buffer (100 mM TrisHCl, 15 mM MgCl2, 500 mM KCl, pH 8.3; Roche), 0.4 µM of each primer (Atr15065LCytB (5′-TAC TCT CCC GAT TCT TTG CCT TCC-3′) and Atr15613H-CytB (5′-AGG ATG AGG AAT AAC GAG AAG TAG AGG A-3′), and approximately 10 ng of DNA template, adjusted to a final volume of 50 µl with PCR-grade H2O.
The optimised cycling conditions for the COI region were a modified version of those of Ward et al. (2005) and consisted of: (i) an initial denaturation phase of 2 minutes at 95˚ C; (ii) 30 amplification cycles, with each cycle consisting of 30 seconds of denaturation at 94° C, 30 seconds of annealing at 54° C, 60 seconds of extension at 72° C; and (iii) a final 10 minute extension at 72° C. 38
Chapter 2. General Methods
Each COI PCR reaction mixture was the same as described for cytochrome b except the use of 0.2 µM of each primer (Fish F1 5′- TCA ACC AAC CAC AAA GAC ATT GGC AC -3′ and Fish R1 5′- TAG ACT TCT GGG TGG CCA AAG AAT CA -3′).
The presence (or otherwise) of PCR product for each reaction was assessed using electrophoresis on a 2% agarose gel, stained with ethidium bromide or SYBR®Safe.
2.4.2
Sequencing
Prior to sequencing, PCR products were cleaned using a QIAquick® PCR Purification Kit (Qiagen 2006). A mixture of 40 µl of a PCR product and 200 µl of Buffer PBI was centrifuged in a spin column for one minute at >17,000 x g. The product was washed with 0.75 ml of Buffer PE and centrifuged for one minute at >17,000 x g. To elute the DNA, 30 µl of Buffer EB was added, and the mixture was left to stand for one minute and then centrifuged at >17,000 x g for one minute.
Approximately 30 ng of clean PCR product was added to 1 µl of PCR grade water, 1 µl (3.2 pmol) of either the forward or reverse primer and 4 µl of Big Dye dye-terminator cycle sequencing ready reaction kit (Applied Biosystems). The amplification conditions were as follows: (i) an initial denaturation phase of 2 minutes at 96˚ C; and (ii) 25 amplification cycles, with each cycle consisting of 10 seconds of denaturation at 96° C, 5 seconds of annealing at 50° C and 4 minutes of extension at 60° C.
39
Chapter 2. General Methods The amplification product was precipitated by adding 1 µl of 3 M sodium acetate (pH 4.6), 1 µl of 125 mM EDTA and 25 µl of 100% ethanol and allowing to stand for 20 minutes before centrifuging for 30 minutes at >17,000 x g and discarding the supernatant. The precipitate was washed with 125 µl of 70% ethanol, centrifuged for 10 minutes at >17,000 x g and air-dried at room temperature.
Sequences were determined using an automated sequencer (ABI Prism 3730) and ABI Prism DNA sequencing analysis software. Heavy and light strands were sequenced for each individual and a consensus sequence generated using the software GENETOOL Lite version 1 (BioTools Inc.).
2.4.3
Trials involving other mtDNA markers
In addition to cytochrome b and COI, attempts were made to PCR-amplify two other regions in the mtDNA - the 16S rRNA gene region and the control region. The attempts were successful for 16S but preliminary data indicated that this gene had the lowest level of variability among the assayed regions and hence was probably less informative for addressing the questions in this thesis, which are concerned with intra-specific variation and variation among closely related species (see Johns and Avise 1998).
Since it is typically the most variable region of the mtDNA (Avise 2000), the control region was initially expected to be the most useful part of the mtDNA for resolving questions about the population structures and phylogeography of Arripis species. Accordingly, attempts were made to amplify this region from each of A. trutta, A. truttaceus and A. georgianus. However, despite a variety of different approaches, the 40
Chapter 2. General Methods control region could not be reliably amplified from these species, possibly due to complications from the presence of a large repeat motif. Full details are provided in Appendix A.1.
2.5
Nuclear DNA
The analysis of nuclear DNA variation in each species of Arripis was based on patterns of fragment length variation at two to four polymorphic EPIC (exon-primed introncrossing) amplified loci, depending on the species. The loci were Aldolase B [intron 4] AldoB4; Calmodulin [intron 4] CaM4; S7 ribosomal protein gene [intron 2] S72; and Creatin kinase CK7 (Appendix A.2, Chapter 3). These loci were chosen from a pool of 28 potential loci (Appendix A.2) because they amplified and exhibited polymorphism in at least one species of Arripis (see Section 2.5.3). Data from an additional eight loci, which were monomorphic, at least within each species, were also used in the assessment of the levels of genetic variation within each species and within the family (Appendix A.2; Chapters 4, 5, 6).
2.5.1
PCR Amplifications
PCR reactions were carried out in a GeneAmp 9700 Thermal Cycler (Applied Biosystems), with: (i) an initial denaturation of 5 minutes at 94˚ C; (ii) 38 amplification cycles, with each cycle consisting of 30 seconds of denaturation at 94° C, 30 seconds of annealing at optimised temperatures (Table 2.2), 30 seconds of extension at 72° C; and (iii) a final 20 minute extension at 72° C (see Section 2.5.4). The only exceptions were: (i) the locus CK7 in A. georgianus, which was amplified using a touch-down PCR of 57 41
Chapter 2. General Methods amplification cycles, with annealing starting at 58° C for 60 seconds and decreasing by 1° C and 1 second each cycle; and (ii) the locus CaM4 in all species, which was amplified using a touch-down PCR of 50 amplification cycles, with annealing starting at 65° C for 60 seconds and decreasing by 0.4° C each cycle (Table 2.2). The PCR reaction mixture contained 0.1 mM of each dNTP (Promega), 0.25 U Taq DNA polymerase (Roche), 1.5 µl Taq Buffer (100 mM Tris-HCl, 15 mM MgCl2, 500 mM KCl, pH 8.3; Roche), 0.04-0.06 µM of each primer (Table 2.2), and approximately 10 ng of DNA template, adjusted to a volume of 15 µl with PCR grade H2O. Forward primers were fluorescently labelled with 6-FAM (CaM4), VIC (CK7) or NED (AldoB4, S72).
2.5.2
Fragment Analysis
PCR products were assayed either singularly or multiplexed for one large and one small locus, using optimised volumes for each locus (Table 2.2).
The products were
precipitated using 1 µl of 3 M sodium acetate (pH 4.6), 1 µl of 125 mM EDTA and 25 µl of 100% ethanol and allowed to stand at room temperature for 13 minutes before centrifuging for 30 minutes at >17,000 x g and discarding the supernatant.
The
precipitate was washed with 125 µl of 70% ethanol, centrifuged for 10 minutes at >17,000 x g and air dried at room temperature.
Following resuspension in 15 µl of formamide, each PCR product was loaded into a 96well plate and fragment length was determined with an ABI 3730 Automated Sequencer (Applied Biosystems) using 0.09 µl of LIZ1200 (Applied Biosystems) size standard per
42
Chapter 2. General Methods well. Allele peaks were manually scored to one decimal place using the software GENEMARKER version 1.71 (SoftGenetics Inc.).
To the best of my knowledge, this is the first study to use capillary electrophoresis to detect and score fragment length variation at EPIC amplified intron loci; previously only agarose or denaturing polyacrylamide gels have been used for such. Capillary electrophoresis, and associated genotyping software, was used on the expectation that it would offer a superior resolution to agarose (Giribet and Wheeler 1999; Sánchez-Pérez et al. 2006) and is faster and less toxic than polyacrylamide (Sánchez-Pérez et al. 2006).
43
Table 2.2 PCR and plate loading conditions for nuclear intron loci assays in the four species of Arripis. In each case, the optimised annealing temperature and primer concentration used in PCR reactions is given, along with the volume of PCR product loaded onto plates for fragment analysis.
A. georgianus Anneal. Primer temp. conc. (µM) (°C)
Load vol. (µl)
A. trutta Anneal. temp. (°C)
Primer conc. (µM)
AldoB4
57
0.04
0.5
56
CaM4
65-45*
0.04
1.0
CK7
58-41*
0.04
S72
64
0.06
Locus
* Touch-down PCR – see text.
Load vol. (µl)
A. truttaceus Anneal. Primer temp. conc. (µM) (°C)
Load vol. (µl)
A. xylabion Anneal. Primer temp. conc. (µM) (°C)
Load vol. (µl)
0.04
2.0
56
0.04
2.0
56
0.04
2.0
65-45*
0.04
1.0
65-45*
0.04
1.0
65-45*
0.04
1.0
2.0
53
0.06
3.5
54
0.06
2.0
53
0.06
3.5
2.5
62
0.06
2.0
63
0.06
2.5
62
0.06
2.0
Chapter 2. General Methods 2.5.3
Primer Selection and Optimisation
As noted earlier, the EPIC primers were selected from a pool of published primers. There were three main phases to this selection process, as described below.
Firstly, 28 sets of published EPIC primers were selected from the literature with preference given to loci with reported length polymorphism, cross-species amplification and reliability (Appendix A.2). Primers were taken from studies including crustaceans, marine and freshwater fishes and cetaceans (see Appendix A.2).
Secondly, preliminary trials showed that 13 of the 28 sets of published primer could be used to amplify a product in representatives of each species of Arripis using touchdown PCR conditions and these 13 loci were therefore selected for further testing. These tests trialled various conditions to determine, for each of the 13 loci, the optimum PCR conditions, including: (i) cycle number; (ii) annealing temperature; (iii) extension times; and (iv) primer concentration.
Thirdly, following the screening of ten individuals of each species, four of the 13 loci (AldoB4, CaM4, S72 and CK7) were found to be polymorphic in at least one species (Appendix A.2) and were screened for length polymorphism as described above. Some further development of the assays conditions for each of these four loci was required to determine the optimal volume of PCR product to be loaded onto the plates for fragment length analysis (Table 2.2).
45
Chapter 2. General Methods 2.5.4
Allele Scoring
Three methods were employed to reduce the potential for errors in scoring genotypes at the intron loci.
Firstly, both positive controls (assays for individuals whose genotype was already known) and negative controls (PCR assays with no added DNA template) were included on most fragment analysis plates. These provided a means for detecting, respectively, inconsistencies in allele scoring across assays (see below) and the presence of any foreign DNA in the assays.
Secondly, a long final extension time (20 minutes) was used to PCR-amplify the alleles at all loci. This was done to maximise the potential of incorporating a “poly-A” on all amplified PCR products and thereby make it easier to distinguish between real alleles and those with an added poly-A (see Figure 2.2).
NB: Non-templated nucleotide
additions by the Taq DNA polymerase (usually a single adenosine to the 3’ end of the fluorescently labelled strand; e.g. Smith et al. 1995; Anon. 2004) are well known in PCR-based studies particularly when short extension times are used (Clark 1988).
Thirdly, all genotypes were scored twice (blindly). Where a discrepancy was detected, the chromatogram was re-examined, and if the score remained ambiguous, the individual was re-assayed by independent PCR and genotyping.
46
Chapter 2. General Methods
0 minutes
5 minutes
15 minutes
20 minutes
Figure 2.2
Examples of the effect of final extension time (0, 5, 15, 20 minutes) on
poly-A addition to the 177 bp allele of AldoB4 in Arripis georgianus. The x-axis represents fragment length in bp. Longer extension time resulted in a greater proportion of alleles exhibiting the poly-A extension as indicated by the increasing height of the second peak. See text for detailed explanation.
47
Chapter 2. General Methods Throughout this study, allele scoring at the intron loci was based on a conservative allele binning approach such that all scores within 20 decimals (where 10 decimals = 1 bp) of another for each of the two larger loci, CK7 (maximum allele size = 984 bp) and S72 (maximum allele size = 931 bp), and within 10 decimals for each of the smaller loci, AldoB4 (maximum allele size = 191 bp) and CaM4 (maximum allele size = 504 bp), were pooled into a single allele class (Figures 2.3 to 2.6, which for simplicity present the data for each locus for only a single species, but the details were similar for the other species). Each allele class (bin) was named by the modal integer and bins were consistent for each locus across all species. This is believed to have minimised the chance of scoring PCR or fragment analysis artefacts of the same allele as a different allele (see below). Although it is possible that this method sometimes resulted in the pooling of different alleles into the same allele class, this type of error is likely to have less significant implications for the data interpretation than the reverse, i.e. it will lead to an underestimate of the extent of genetic diversity and population structure rather
178
45 40 35 30 25 20 15 10 5 0
191
194
193
192
191
190
189
188
187
186
185
185
184
183
182
181
180
179
178
177
176
175
182
174
173
172
172
171
170
170
Frequency
than an overestimate.
Fragment Length (bp)
Figure 2.3
Frequency histogram of fragment sizes for the locus AldoB4 in all
assayed Arripis trutta. Bars and arrows indicate allele bins, labelled with the modal integer (see text).
48
Chapter 2. General Methods
491
486 499 470
480
483
504
468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505
Frequency
50 45 40 35 30 25 20 15 10 5 0
Fragment Length (bp)
Frequency histogram of fragment sizes for the locus CaM4 in all assayed
Figure 2.4
Arripis georgianus. Bars and arrows indicate allele bins, labelled with the modal integer (see text).
Freq u en cy
952
18 16 14 12 10
961 943
984
8 6 4
940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984
2 0 Fragment Length (bp)
Figure 2.5
Frequency histogram of fragment sizes for the locus CK7 in all assayed
Arripis trutta. Bars and arrows indicate allele bins, labelled with the modal integer (see text).
49
Chapter 2. General Methods
914
16 14
919
Frequency
12
900
10 8
892
924
931
895
6 4 2
890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932
0
Fragment Length (bp)
Figure 2.6
Partial frequency histogram of fragment sizes for the locus S72 in all
assayed Arripis trutta. Bars and arrows indicate allele bins, labelled with the modal integer (see text). Note that for scaling reasons, the x-axis has been abbreviated at the origin and thus a single occurrence of each of two alleles (816 & 869) is not shown.
Allele binning, which is common place in studies of fragment length variation at intron loci (e.g. Gomulski et al. 1998; Berrebi et al. 2005; Hubert et al. 2006; Atarhouch et al. 2007), was necessary because of variability among repeats of the same individual (Table 2.3). Bins encompassing 20 decimals were used for large loci (CK7 and S72), because repeat assays indicated intra-individual variability may be up to 1.5 bp and the discrepancies would therefore sometimes have been scored as a different allele if allele bins had not been used (Table 2.3). Although repeat assays at the smaller loci (AldoB4 and CaM4) indicated that mis-scoring due to intra-individual variability was unlikely, there was still variability – less than 0.5 bp (Table 2.3). Consequently, alleles within 10 decimals were binned in these small loci. Variability among repeat assays may be due to PCR and/or electrophoretic factors (see Appendix A.3).
50
Chapter 2. General Methods Binning of fragment length alleles at EPIC amplified intron loci is usually based on a subjective approach with alleles placed into arbitrary, fixed size classes, usually five (e.g. Hubert et al. 2006; Atarhouch et al. 2007), ten (e.g. Berrebi et al. 2005) or more (e.g. Gomulski et al. 1998) bases in size. Sometimes bins are required due to the inaccuracies of ‘by eye’ scoring of gels (e.g. Atarhouch et al. 2007), whereas sometimes binning is used at the post-scoring stage to address ambiguities in the data (e.g. Hubert et al. 2006). The method of allele binning in this study, which is based on gaps in the frequency distribution of the scoring of the raw alleles, is thought to be more meaningful than the arbitrary allele size classes used in previous studies (see also Appendix A.3).
Table 2.3
(over page) Results of tests of variability in raw scoring (i.e. without
binning) of intron loci allele lengths in repeat assays in Arripis. The tests were based on (two) repeat assays for n individuals. Some tests were done using different PCR reactions, with the products loaded onto the same or different fragment analysis plates, whereas other tests were done using the same PCR reaction, with the product loaded onto the same or different plates. Mean, standard error and maximum errors (given as integer) are the difference between scores (in bp) for the two repeats. The percentage of repeats where a different allele would have been scored if allele bins had not been used is given as ‘% different’. For AldoB4, CK7 and S72, data are pooled across all alleles for all three Australian species. In CaM4, only data for A. georgianus were used (because this was the only species for which this locus was polymorphic).
51
Chapter 2. General Methods
Allele AldoB4 n
Different PCR Same plate Different plate 1 2 1 2
Same PCR Same plate Different plate 1 2 1 2
(maximum allele size = 191 bp) 12
12
39
39
64
64
6
6
0.03
0.05
0.1
0.09
0.06
0.05
0.05
0.06
± 0.02
± 0.02
± 0.02
± 0.02
± 0.01
± 0.01
± 0.01
± 0.01
maximum
0.2
0.2
0.4
0.4
0.2
0.2
0.2
0.2
% different
0
0
0
0
0
0
0
0
mean s.e.
(maximum allele size = 504 bp)
CaM4 n
6
6
22
22
6
6
2
2
0.03
0.1
0.1
0.11
0.07
0.07
0.1
0.1
± 0.03
± 0.03
± 0.02
± 0.02
± 0.04
± 0.02
± 0.00
± 0.00
maximum
0.2
0.2
0.2
0.2
0.2
0.1
0.1
0.1
% different
0
0
0
0
0
0
0
0
mean s.e.
(maximum allele size = 984 bp)
CK7 n
49
49
75
75
47
47
9
9
0.27
0.32
0.41
0.47
0.24
0.27
0.29
0.29
± 0.02
± 0.03
± 0.04
± 0.04
± 0.02
± 0.02
± 0.08
± 0.08
maximum
0.9
1.5
1.2
1.5
0.9
1.2
0.6
0.6
% different
25.3
24.2
36
42.7
28.4
21.1
44.4
22.2
mean s.e.
(maximum allele size = 931 bp)
S72 n
45
45
26
26
40
40
21
21
0.15
0.22
0.23
0.29
0.22
0.26
0.21
0.23
± 0.01
± 0.02
± 0.03
± 0.04
± 0.02
± 0.03
± 0.04
± 0.04
maximum
0.5
1
0.5
0.9
0.9
1.6
0.6
0.8
% different
10
7.8
26.9
26.9
18.8
18.8
9.5
14.3
mean s.e.
2.6
Nuclear DNA Sequences
An attempt was made to determine the nucleotide sequence of some EPIC markers in order to provide additional sequence data (i.e. in addition to the mtDNA data), as has been successfully conducted in other similar studies (e.g. Lessa 1992; Hare et al. 2002; Lavoué et al. 2003; Teske et al. 2004; Morrison et al. 2006). This attempt was based on simultaneous sequencing of both of the alleles at a diploid locus and recognising 52
Chapter 2. General Methods heterozygotes as two peaks of nearly equal intensity at the same base position in the chromatograms (Creer 2007). These sites may then be scored using standard IUPAC ambiguity codes. When successful, this method avoids the time and costs associated with the cloning and multiple sequencing required to independently ascertain the sequence of both alleles (Lavoué et al. 2003; Teske et al. 2004; Morrison et al. 2006), although the method has some potential disadvantages (e.g. Hare 2001; Creer 2007). The most significant problem arises when heterozygotes contain alleles of different length.
Insertions and deletions (indels) within alleles will result in unreadable
chromatograms as the two superimposed sequences will be out of phase and distinguishing between them becomes impossible (Mallarino et al. 2005; Creer 2007).
The intron loci trialled for direct sequencing were those loci that could be amplified in Arripis and for which no length variants were detected (AldoB1, Am2B-1, GnRH3-1, GnRH3-2, GnRH3-3, Gpd2, Mlc-2-c, S71; see Appendix A.2). PCR was first used to amplify DNA product with conditions previously optimised for each gene (data not shown). Of the eight loci tested, four gave indecipherable chromatograms, which were probably caused by numerous indels, the sum of which maintained length monomorphism among alleles, but still caused the sequences to be out of phase. The other four loci (AldoB1, GnRH3-1, GnRH3-2, GnRH3-3) were then amplified using the same touch-down PCR conditions described for CaM4 in Section 2.5.1. Each PCR reaction mixture was the same as that described in Section 2.4.1 (see Appendix A.2 for primer sequences), and cleaning, sequencing and alignment was as described in Section 2.4.2. However, the amount of sequence polymorphism at each of the four loci was extremely low and provided little utility for the purposes of this thesis.
53
Chapter 3. Population Genetic Structure
3.0
3.1
Population Genetic Structure
Introduction
As outlined in Chapter 1, the primary goal of this thesis was to examine the distribution of genetic variation within and between Arripis species and to use the resultant information to infer aspects of the origins and evolutionary histories of these species. However, before using this information to address evolutionary questions, it is important to have an understanding of the population genetic structure of each species and associated patterns of contemporary gene flow. This is important because the extent and patterns of contemporary gene flow will have a major bearing on the distribution of genetic variation within each species. As a part of this, an understanding of the patterns of contemporary gene flow can also be used to identify the most appropriate study units (e.g. genetically differentiated populations or the entire species) for the evolutionary questions. Hence, this chapter uses length variation among alleles at nuclear loci to investigate the population genetic structure of each of A. trutta, A. truttaceus and A. georgianus in Australian waters.
Regardless of the thesis-wide
perspective, a study of the population genetic structures of these species is significant in its own right in the context of the current interest in the extent, patterns and determinants of contemporary gene flow in marine systems (e.g. Andre´ et al. 2011; Toonen et al. 2011).
As described in Chapter 1, tagging and otolith microchemistry studies indicate that each of A. trutta, A. truttaceus and A. georgianus have complex life histories involving a very high dispersal potential (see Figure 1.3). Specifically, in all species: (i) pelagic 54
Chapter 3. Population Genetic Structure larvae are passively transported over potentially vast distances; (ii) post-settlement and sub-adult fish progressively return to the respective adult distributions over several years; and (iii) adults of all species make seasonal and spawning movements within their adult distributions (see species accounts in Chapter 1). In fact, based on this lifehistory, few coastal marine fishes appear to share the dispersal potential of Arripis (Ward et al. 1994; Gillanders et al. 2003).
On the basis of possessing a seemingly high dispersal potential at every life stage, it is tempting to conclude that the species of Arripis provide some rare examples of coastal pelagic species that are panmictic (random mating) over their entire range (e.g. Waples 1987; Ward et al. 1994; Siegel et al. 2003). However, although, in the past, panmixia has been assumed for many coastal pelagic fishes, most studies of species with a seemingly highly dispersive life-style have, in fact, detected the presence of population genetic structure (e.g. Gold et al. 2002; Knutsen et al. 2003; Rohfritsch and Borsa 2005; Ruzzante et al. 2006a; Atarhouch et al. 2007; Chairi et al. 2007; Marko et al. 2007; Chlaida et al. 2008; Wang et al. 2008, but see Hoolihan et al. 2006; Cárdenas et al. 2009; Palm et al. 2009).
This emphasises the importance of compiling detailed
information before jumping to conclusions. Testing the assumption of panmixia is especially important in exploited species, such as Arripis, for which fisheriesmanagement decisions are (or should be) based on an understanding of connectivity and dispersal (Sale et al. 2005; Ruzzante et al. 2006a).
As established in Chapter 1, there is a need for analysis of the population genetic structure of Arripis species based on DNA markers. Certainly, previous genetic studies on these species have been based on allozyme markers with extremely low levels of 55
Chapter 3. Population Genetic Structure polymorphism (MacDonald 1980; Ayvazian et al. 2004). In addition, only the genetic results for A. georgianus have been published (Ayvazian et al. 2004).
3.1.1
Aims and Hypothesis
Length polymorphism in several nuclear introns was used to examine the genetic structure of each of A. trutta, A. truttaceus and A. georgianus in Australian waters, and test the following hypothesis.
Each of the three Arripis species is genetically homogenous across its Australian distribution.
This is predicted on the basis that each of these species undergoes extensive movements at all stages of life and the potential for mixing is high.
The reasons for using length polymorphism in introns for this analysis are explained in Section 1.3.1.2. Ideally, it would have been useful to complement the nDNA data with information from a mtDNA marker (e.g. Chow and Takeyama 2000; Haney et al. 2007), however, while mtDNA sequence data were collected for these species, the levels of polymorphism were not adequate to test for population structure (see Chapters 4, 5 and 6) and it was not possible to reliably amplify the potentially more variable control region (see Chapter 2).
56
Chapter 3. Population Genetic Structure 3.2
Materials and Methods
3.2.1
Sampling
The results of this chapter are based upon the analysis of samples of: (i) A. trutta from three sampling sites (NSW, Victoria, Tasmania); (ii) A. truttaceus from five sampling sites ranging from Augusta, Western Australia to The Coorong, South Australia; and (iii) A. georgianus from seven sampling sites ranging from Fremantle, Western Australia to The Coorong, South Australia (see Chapter 2, Figure 2.1 for details). For A. trutta, the number of individuals per sample ranged from 29 to 30 (total number of individuals = 89), while the range for A. truttaceus was 9 to 39 (total number of individuals = 130) and that for A. georgianus was 9 to 44 (total number of individuals = 220) (see Chapter 2 for details). While most sites were only sampled once, a sample of A. georgianus was collected from one site (Fremantle) once in 2006 and again in 2008 in order to test for temporal changes at a site. Individuals of A. georgianus were measured for total length in order to provide an estimation of age or if only a head was available, total length was estimated from the relationship between head length and total length. Most samples included fish of various sizes and thus contained fish of different ages (i.e. multiple cohorts).
3.2.2
Genetic Assays
The results of this chapter are based on the analysis of length polymorphism from two to four EPIC amplified intron loci per species, depending on the species. In particular, the results for A. trutta are based on length variation at the loci AldoB4, S72, CK7; those 57
Chapter 3. Population Genetic Structure for A. truttaceus are based on the loci AldoB4, S72; and those for A. georgianus are based on the loci AldoB4, CaM4, S72, CK7 (see Chapter 2). The target loci for each species were selected for analysis on the basis that they could be reliably amplified from that species and that they exhibited scorable length polymorphism, where the frequency of the most common allele was less than 0.95 in pooled samples for a species (see Chapter 2 for details of locus selection). The alleles at each locus were PCR-amplified and screened and scored using the methods described in Chapter 2. The alleles at each locus were binned into size classes as explained in Chapter 2 (Section 2.5.4). The frequencies of (binned) alleles were calculated from the raw genotype data.
3.2.3
Statistical Analyses
3.2.3.1
Levels of Polymorphism
For each species, the number of alleles per locus (NA), observed heterozygosity (HO) and expected unbiased gene diversity (expected heterozygosity, HE, Nei 1987) were calculated for each sample and over all samples.
3.2.3.2
HWE, Linkage Disequilibrium and Cohort Effects
Exact tests (Guo and Thompson 1992) were used to assess the statistical significance of differences between the observed numbers of homozygotes and heterozygotes and the numbers expected under Hardy-Weinberg Equilibrium (HWE) conditions at each intron locus in each sample of Arripis. This was done to test for the existence of patterns of non-random mating (see Nei 1987) in the sampled assemblages of Arripis and/or the 58
Chapter 3. Population Genetic Structure presence of PCR artefacts such as null alleles and large allele drop-out in the data set (see Pompanon et al. 2005). The exact tests used the Markov chain method to estimate the exact probability of a Type I error, as implemented by GENEPOP version 4.0 (Raymond and Rousset 1995a). The iteration parameters for these, and all subsequent tests conducted with GENEPOP, were 10,000 dememorisation steps, 1,000 batches and 10,000 iterations per batch. Exact probability tests are not considered to be biased by small sample sizes or low frequencies of alleles or genotypes (Raymond and Rousset 1995a) and so are appropriate for the EPIC data of this research.
A sequential
Bonferroni correction (Rice 1989) was applied to control for group-wide Type I error.
Exact probability tests were also used to test for evidence of non-random associations of genotypes between loci (i.e. for departures from linkage disequilibrium) in the samples of Arripis, using GENEPOP. Exact probability tests were conducted for all pairwise combinations of loci in each sample of each species. A global test for all samples within a species was also conducted using an exact chi-square test with Fisher’s combined probability method in GENEPOP (Raymond and Rousset 1995b).
Since the presence of multiple age-class cohorts could potentially affect the results through events such as annual variations in recruitment (e.g. Johnson and Black 1984; Hedgecock 1994), a limited test for the effects of cohort on genetic variation was conducted using the A. georgianus data. This was done by assigning individuals of A. georgianus to one of four year classes using the length-age relationship developed for this species by Fairclough et al. (2000b) (Table 3.1) and then using a hierarchical Analysis of Molecular Variance (AMOVA) to investigate how genetic variation was partitioned within and among the cohorts from the different collection sites.
The 59
Chapter 3. Population Genetic Structure AMOVA was conducted using ARLEQUIN version 3.1 (Excoffier et al. 2005), with statistical significance determined by a non-parametric approach over 10,000 permutations. Arripis georgianus was selected for this analysis because the sampled individuals of this species exhibited the widest range of sizes and a calibrated method of estimating age groups of individuals was available (Fairclough et al. 2000b).
Table 3.1
Sample size for each cohort within each spatial sample of Arripis
georgianus. Individuals were assigned to cohorts based on the length-age relationship of Fairclough et al. (2000b). Collection sites are ordered from the most northerly to the most south-easterly (see Table 2.1 for details).
Cohort 2003 2004 2005 2006
HF 11 17 2 0
HC 3 11 21 0
HP 3 30 1 0
HI 6 17 0 0
HO 8 16 1 0
HW 5 0 1 0
HS 0 0 4 31
On the basis that the hierarchical AMOVA indicated that there was no cohort effect (see Results, Section 3.3.2), the rest of the analyses for each species proceeded using pooled cohorts. Note that the age-based nature of migration in these fishes means that it is difficult to sample all cohorts at a single site (see Table 3.1), and as a consequence the AMOVA (for A. georgianus) unavoidably had many low cell values, which might reduce its power, although the p-values were far from significant.
60
Chapter 3. Population Genetic Structure 3.2.3.3
Population Differentiation
Several approaches were used to assess population genetic structure in each of A. trutta, A. truttaceus and A. georgianus, as detailed below. The 2008 sample of A. georgianus was not included in any of the spatial analyses, except as described in Section 3.2.3.8.
3.2.3.4
Pairwise Differentiation
FST, the standardised genetic variance, was used as a measure of the proportion of the total allele frequency variation due to differences between samples for each species of Arripis. If genetic heterogeneity exists among samples, FST values will be significantly greater than zero.
Pairwise FST tests between samples within each species were
conducted with the permutation approach in FSTAT version 2.9.3.2 (Goudet 1995), with 10,000 permutations assuming samples in HWE. Using the method of Weir & Cockerham (1984), FST values represented weighted averages of variation across loci.
There is increasing recognition of the importance of standardising FST according to the level of polymorphism detected within a species in order to compare across species and markers (e.g. Hedrick 2005; Meirmans 2007), but given that the FST values for Arripis were effectively zero or nearly so (see Results), standardisation was not applicable in this case.
The pairwise relationships among the samples, as measured by pairwise values of FST, of each species was summarised in a multi-dimensional scaling (MDS) plot that was constructed with 9,999 restarts in PRIMER version 5.2.9 (Clarke and Gorley 2002). 61
Chapter 3. Population Genetic Structure
3.2.3.5
Overall Differentiation
Single-locus FST estimates were made using FSTAT, as described in Section 3.2.3.4. The statistical significance of differences in the allele frequencies at each locus between all samples of a species was determined as the probability of obtaining, by chance, a value as large as or larger than the observed FST.
Differentiation over all loci between spatial samples was tested using a re-sampling permutation test utilising 10,000 permutations with the software GENETIX version 4.05.2 (Belkhir et al. 2004) and generating a probability estimate by counting the number of times a randomly permuted FST ≥ actual FST.
3.2.3.6
Factorial Correspondence Analysis
The factorial correspondence analysis (FCA) of GENETIX was used to search for structure in the data. In this analysis, composite axes were constructed using allelic data so that inter-individual differences were maximised and the result visualised by plotting onto a two-dimensional chart. Guinand (1996) showed that the values of inertia along each axis can be regarded as combinations of single locus FST linear values, thus FCA is a valid method to examine heterogeneity using FST. It differs from other methods by using individuals (rather than sub-populations) as the unit of analysis.
62
Chapter 3. Population Genetic Structure
3.2.3.7
Power Analysis
A retrospective power analysis was conducted with 1000 runs using the method described by Theisen et al. (2008) and the software POWSIM version 4.0 (Ryman and Palm 2006). The actual data (i.e. number of populations and their sample sizes, number of alleles at each locus and their observed frequencies) were used to generate 1000 random populations that were then allowed to drift to various levels of genetic divergence (quantified by FST and determined by a combination of effective population size, Ne, and generations of drift, t). These simulated populations were sampled, and tested against the null hypothesis of genetic homogeneity among populations using a chi-square approach and Fisher’s exact test. The percentage of significant outcomes (p < 0.05) from this resampling was interpreted as the power of the data to detect the defined level of genetic divergence. For each simulation in POWSIM, t was altered, while Ne was held constant at 1000. The simulations were repeated until the lowest level of genetic divergence (i.e. lowest value of t) that could be detected among the sampled populations of each species with a power ≥ 95% was identified. The software cannot accept different sample sizes for different loci within a sample, so the analysis used a reduced dataset by removing individuals with missing data. Therefore, this method provides an underestimate of the true power of the data. Sample sizes for the power analysis were as follows: (see Table 2.1 for sample codes): A. trutta EN (28), EV (27), ET (27); A. truttaceus WA (27), WY (15), WL (38), WW (8), WS (31); A. georgianus HF (30), HC (35) HP (36), HI (21), HO (25), HW (6), HS (35).
63
Chapter 3. Population Genetic Structure
3.2.3.8
Temporal Stability
Extensive temporal sampling was not undertaken, partly because preliminary analyses revealed no evidence of significant spatial genetic heterogeneity in any of the three study species. Nevertheless, two samples of A. georgianus collected from the same site (Fremantle, Western Australia) two years apart were compared, providing a very limited test of the extent of any temporal heterogeneity. Three approaches were used to test for temporal stability.
Firstly, the values of FST between the sample collected from
Fremantle in 2008 and each of the seven samples of A. georgianus, including one from Fremantle, collected in 2006 were plotted in a MDS, as described above. Secondly, an AMOVA between just the two temporal samples (HF and HF08) was conducted. By nesting the two samples into a single group, this design directly tests for temporal differences (e.g. Hale et al. 2001). Thirdly, the A. georgianus samples collected from Fremantle in 2008 were added to the FCA, and graphed as described above.
3.3
Results
3.3.1
Polymorphism, HWE and Linkage Disequilibrium
Length polymorphism at four EPIC amplified intron loci (AldoB4, S72, CK7, CaM4) was used to investigate the population genetic structure of A. georgianus. The first three of these loci also exhibited length polymorphism in the samples of A. trutta and so were also used to investigate the population genetic structure of this species. Similarly, the first two loci exhibited length polymorphism in the samples of A. truttaceus and 64
Chapter 3. Population Genetic Structure were used for this species. Based on all samples for all species combined, the alleles at the locus AldoB4 ranged from 170 to 191 bp, while those at the locus CaM4 ranged from 469 - 504 bp and those of S72 and CK7 ranged between 816 – 931 bp and 844 984 bp, respectively (Table 3.2).
The overall level of polymorphism was greatest in the samples of A. georgianus (four polymorphic loci, with a total of 32 different alleles), followed by those of A. trutta (three polymorphic loci, with a total of 20 different alleles), while those of A. truttaceus contained only two polymorphic loci, with 9 different alleles (Table 3.2).
Mean
expected heterozygosity per sample per locus, when the locus was polymorphic in the sample, was typically low to moderate, for example it ranged from 0.083 (CaM4 locus in the HI sample of A. georgianus) to 0.655 (S72 locus in the EN sample of A. trutta) (Table 3.2).
There were no statistically significant, non-random associations of genotypes at different loci for the samples of each of the three species (once a Bonferroni correction was applied); thus the patterns of variation at the different intron loci are assumed to be independent of each other (Tables 3.3, 3.4, 3.5). The genotype frequencies at each locus in each sample of each of the three species were generally in accordance with those expected under HWE. Specifically, of 49 tests, only a single departure from HWE was detected (CK7 locus in the HO sample of A. georgianus; Table 3.2) and this was not statistically significant following Bonferroni correction.
65
Chapter 3. Population Genetic Structure
Table 3.2
(over page)
Allele frequencies at each of four intron loci for samples
of n individuals of three species of Arripis. NA is the number of alleles for each sample and for the species (in parentheses).
HE and HO are expected and observed
heterozygosity, respectively. HWE(p) is the probability of a Type 1 error in exact tests for departures from Hardy-Weinberg equilibrium expectations. A dash (-) indicates the allele was not recorded. Collection sites are ordered from the most northerly to the most south-easterly (see Table 2.1 for details).
66
Arripis trutta EN EV ET AldoB4 locus n 29 5 NA
30 3
30 3
WA
(6)
Arripis truttaceus WY WL WW
Arripis georgianus HI HO HW
WS
HF
HC
HP
44 5
39 2
36 4
23 4
28 3
30 2
15 2
39 2
9 2
37 2
(2)
HS
HF08
9 3
41 4
41 5
170 172 175 178 182 185 191
0.000 0.017 0.741 0.034 0.034 0.172
0.000 0.017 0.783 0.000 0.000 0.200
0.017 0.000 0.800 0.000 0.000 0.183
0.600 0.400 -
0.733 0.267 -
0.615 0.385 -
0.778 0.222 -
0.743 0.257 -
0.114 0.011 0.023 0.818 0.034 -
0.064 0.000 0.000 0.936 0.000 -
0.153 0.000 0.014 0.819 0.014 -
0.152 0.000 0.043 0.783 0.022 -
0.089 0.018 0.000 0.893 0.000 -
0.056 0.000 0.000 0.833 0.111 -
0.061 0.012 0.012 0.915 0.000 -
0.073 0.024 0.000 0.902 0.000 -
HE HO HWE(p)
0.419 0.414 0.33
0.347 0.433 0.45
0.326 0.333 1.00
0.480 0.667 0.06
0.391 0.400 1.00
0.474 0.513 0.74
0.345 0.222 0.34
0.382 0.405 1.00
0.316 0.295 0.50
0.120 0.128 1.00
0.305 0.250 0.43
0.361 0.348 0.71
0.194 0.214 1.00
0.291 0.333 1.00
0.159 0.171 1.00
0.180 0.171 0.16
3 1
3 1
4 1
2 1
2 1
2 1
2 1
2 1
42 5
39 5
36 3
23 3
28 3
8 2
40 6
41 5
1.000 -
1.000 -
1.000 -
1.000 -
1.000 -
1.000 -
1.000 -
1.000 -
0.000 0.000 0.000 0.012 0.060 0.905 0.012 0.012
0.013 0.000 0.013 0.013 0.038 0.923 0.000 0.000
0.000 0.000 0.000 0.000 0.042 0.944 0.000 0.014
0.000 0.000 0.000 0.022 0.022 0.957 0.000 0.000
0.018 0.000 0.000 0.000 0.107 0.875 0.000 0.000
0.000 0.000 0.000 0.000 0.063 0.938 0.000 0.000
0.000 0.013 0.013 0.013 0.075 0.875 0.000 0.013
0.012 0.000 0.000 0.012 0.024 0.939 0.012 0.000
-
-
-
-
-
-
-
-
0.177 0.190 1.00
0.146 0.154 1.00
0.107 0.111 1.00
0.083 0.087 1.00
0.223 0.179 0.35
0.116 0.125 -
0.228 0.200 0.48
0.117 0.122 1.00
CaM4 locus n NA 469 477 480 483 486 491 498 504 HE HO HWE(p)
(1)
(1)
(5)
(8)
Arripis trutta EN EV ET CK7 locus n NA
28 3
28 4
29 3
844 848 923 931 936 939 943 948 952 961 976 984
0.089 0.839 0.071 0.000
0.018 0.893 0.071 0.018
HE HO HWE(p)
0.283 0.321 1.00
0.197 0.214 1.00
WA
Arripis truttaceus WY WL WW
Arripis georgianus HI HO HW
WS
HF
HC
HP
42 6
38 7
35 4
23 3
26 2
HS
HF08
8 1
39 6
38 6
8 1
9 1
8 1
12 1
10 1
0.052 0.845 0.103 0.000
1.000 -
1.000 -
1.000 -
1.000 -
1.000 -
0.000 0.012 0.000 0.000 0.012 0.000 0.060 0.060 0.845 0.012 0.000 -
0.000 0.026 0.013 0.000 0.000 0.013 0.013 0.013 0.908 0.000 0.013 -
0.000 0.014 0.000 0.000 0.000 0.000 0.071 0.029 0.886 0.000 0.000 -
0.000 0.022 0.000 0.000 0.000 0.000 0.087 0.000 0.891 0.000 0.000 -
0.000 0.000 0.000 0.000 0.000 0.000 0.077 0.000 0.923 0.000 0.000 -
0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.000 0.000 0.000 -
0.026 0.051 0.000 0.013 0.000 0.000 0.026 0.038 0.846 0.000 0.000 -
0.026 0.053 0.000 0.000 0.013 0.000 0.026 0.053 0.829 0.000 0.000 -
0.273 0.321 1.00
-
-
-
-
-
0.278 0.214 0.20
0.174 0.184 1.00
0.209 0.229 1.00
0.198 0.217 1.00
0.142 0.154 1.00
-
0.279 0.231 0.13
0.306 0.237 0.06
(4)
(1)
(11)
Arripis trutta EN EV ET S72 locus n NA
29 6
25 7
28 4
816 869 871 875 883 889 892 895 900 914 919 924 931
0.017 0.000 0.000 0.000 0.224 0.500 0.207 0.034 0.017
0.000 0.020 0.020 0.020 0.100 0.580 0.240 0.020 0.000
HE HO HWE(p)
0.655 0.552 0.31
0.594 0.480 0.31
WA
Arripis truttaceus WY WL WW
Arripis georgianus HI HO HW
WS
HF
HC
HP
32 6
36 5
34 4
23 3
26 3
HS
HF08
8 3
36 5
36 4
27 4
15 3
38 4
13 4
45 5
0.000 0.000 0.000 0.000 0.107 0.625 0.250 0.018 0.000
0.000 0.111 0.722 0.148 0.019
0.000 0.233 0.600 0.167 0.000
0.000 0.066 0.697 0.211 0.026
0.000 0.038 0.731 0.192 0.038
0.022 0.178 0.600 0.167 0.033
0.016 0.016 0.063 0.000 0.016 0.188 0.703 0.000 -
0.000 0.000 0.111 0.014 0.028 0.125 0.722 0.000 -
0.000 0.000 0.088 0.000 0.015 0.059 0.838 0.000 -
0.000 0.000 0.152 0.000 0.000 0.022 0.826 0.000 -
0.000 0.000 0.096 0.000 0.000 0.058 0.846 0.000 -
0.000 0.000 0.125 0.000 0.000 0.250 0.625 0.000 -
0.000 0.000 0.083 0.000 0.014 0.083 0.806 0.014 -
0.000 0.000 0.097 0.000 0.028 0.083 0.792 0.000 -
0.535 0.679 0.51
0.444 0.333 0.12
0.558 0.467 0.56
0.465 0.474 1.00
0.426 0.462 0.43
0.579 0.667 0.24
0.466 0.406 0.44
0.450 0.528 1.00
0.286 0.324 1.00
0.294 0.261 0.52
0.272 0.154 0.01*
0.531 0.750 0.70
0.336 0.333 0.63
0.356 0.361 0.06
(9)
* not significant after Bonferroni correction.
(5)
(8)
Chapter 3. Population Genetic Structure Results of exact tests for linkage disequilibrium for all pairwise
Table 3.3
combinations of three intron loci for each sample of Arripis trutta and for all of these samples combined. The results for the pairwise comparisons are presented in terms of the probability of Type I error, as determined by exact tests. The results for the global test are exact probabilities from a chi-square test (χ2), with associated degrees of freedom (df), using Fisher’s combined probability method. Sample codes are as per Table 2.1.
samples Locus Pair
global
EN
EV
ET
χ2
df
p-value
CK7
S72
0.19
0.50
0.58
5.833
6
0.44
CK7
AldoB4
0.32
0.52
0.93
3.752
6
0.71
S72
AldoB4
0.77
0.14
0.31
6.842
6
0.34
Table 3.4
Results of exact tests for linkage disequilibrium for all pairwise
combinations of two intron loci for each sample of Arripis truttaceus and for all of these samples combined. The results for the pairwise comparisons are presented in terms of the probability of Type I error, as determined by exact tests. The results for the global test are exact probabilities from a chi-square test (χ2), with associated degrees of freedom (df), using Fisher’s combined probability method. Sample codes are as per Table 2.1.
samples Locus Pair S72
AldoB4
global
WA
WY
WL
WW
WS
χ2
df
p-value
0.80
0.82
0.75
0.11
0.70
6.631
10
0.76
70
Table 3.5
Results of exact tests for linkage disequilibrium for all pairwise combinations of four intron loci for each sample of Arripis
georgianus and for all of these samples combined. The results for the pairwise comparisons are presented in terms of the probability of Type I error as determined by exact tests. The results for the global test are exact probabilities from a chi-square test (χ2), with associated degrees of freedom (df), using Fisher’s combined probability method. Sample codes are as per Table 2.1.
samples Locus Pair
global
HF
HC
HP
HI
HO
HW
HS
HF8
χ
2
df
p-value
CK7
S72
0.04
0.57
0.23
0.65
1.00
-
0.32
0.55
15.079
14
0.37
CK7
AldoB4
0.46
0.22
0.12
0.06
0.60
-
0.73
0.84
16.422
14
0.29
CK7
CaM4
0.56
0.11
0.66
1.00
0.68
-
0.38
0.68
9.855
14
0.77
S72
AldoB4
0.85
0.58
0.49
0.40
0.50
1.00
0.58
0.77
7.707
16
0.96
S72
CaM4
0.20
0.22
1.00
0.52
0.44
0.43
0.57
0.75
12.571
16
0.70
AldoB4
CaM4
0.93
1.00
0.79
0.36
0.77
1.00
0.05
1.00
9.338
16
0.90
Chapter 3. Population Genetic Structure 3.3.2
Cohort Effects
As mentioned in Section 3.2.3.2, genetic variation among cohorts was negligible. The hierarchical AMOVA, which included individuals grouped into cohorts within each spatial sample of A. georgianus, found that almost all of the variance in the data (99.35%) occurred within cohort groups, although FST was effectively zero (FST = 0.006, p = 0.33). Furthermore, variance due to differences between cohorts within a spatial sample (FSC = 0.002) was less than the variance between spatial samples (FCT = 0.004), although both were effectively zero (p = 0.26 and 0.48, respectively).
3.3.3
Population Differentiation
The values of FST between each pair of samples within each of the three study species were invariably low (e.g. effectively zero for A. trutta, FST < 0.018 for A. truttaceus, and FST < 0.030 for A. georgianus) and none were significantly different from zero (Tables 3.6, 3.7, 3.8). For each species, the distribution of samples in a MDS plot, which is based on pairwise FST values, revealed no clear evidence of a relationship between the genetic composition of samples and the location of the source population (Figures 3.1, 3.2, 3.3).
Within each species, single locus estimates of (global) FST were consistently low (Table 3.9), reflecting the fact that the allele frequencies at each locus were generally similar in all of the spatial samples of a particular species (Table 3.2). Accordingly the (global) multi-locus values of FST for each species were also low (e.g. ≤ 0.005) and not significantly different from zero Table 3.9). 72
Chapter 3. Population Genetic Structure
Factorial correspondence analysis revealed little to no clustering of individuals in any species, with clouds of data points from each spatial sample generally overlapping (Figures 3.4, 3.5, 3.6).
In all cases, no spatial patterns were discernable and the
distribution of the data was not strongly explained by either axis (Figures 3.4, 3.5, 3.6).
3.3.4
Power Analysis
Power analyses suggested that genetic divergence would be detected with ≥ 95% confidence, at least at FST ≥ 0.034 for A. trutta (Ne = 1000, t = 70), FST ≥ 0.039 for A. truttaceus (Ne = 1000, t = 80) and FST ≥ 0.009 for A. georgianus (Ne = 1000, t = 20).
3.3.5
Temporal Stability
No pairwise tests involving the temporal samples of A. georgianus from Fremantle were significant (Table 3.8), and the multi-dimensional (MDS) distance between the two temporal samples was equal to or greater than the distances between most of the other spatial samples (Figure 3.3). AMOVA detected no significant difference between the samples of A. georgianus collected from Fremantle two years apart (FST = 0.012, p = 0.12), and the cloud of individuals from the two samples in the FCA generally overlapped (Figure 3.6).
73
Chapter 3. Population Genetic Structure
Table 3.6
Results of tests for differentiation between pairs of samples of Arripis
trutta presented as pairwise FST values (below diagonal) and the probability of obtaining by chance, a value as large or larger than the observed FST (above diagonal, italicised) using a permutation approach. Sample codes are as per Table 2.1.
EN EN EV ET
Figure 3.1
-0.004 -0.000
EV 0.27
ET 0.32 0.97
-0.013
Multi-dimensional scaling (MDS) plot of pairwise FST values between
samples of Arripis trutta, as presented in Table 3.6. Sample codes are as per Table 2.1. The stress value is an inverse measure of the degree of correspondence between the distances among points implied by MDS and the matrix input.
74
Chapter 3. Population Genetic Structure
Table 3.7
Results of tests for differentiation between pairs of samples of Arripis
truttaceus presented as pairwise FST values (below diagonal) and the probability of obtaining by chance, a value as large or larger than the observed FST (above diagonal, italicised) using a permutation approach. Sample codes are as per Table 2.1.
WA WA WY WL WW WS
Figure 3.2
0.009 -0.011 0.008 0.018
WY 0.40 0.012 -0.011 -0.019
WL 0.83 0.11 0.000 0.018
WW 0.49 0.22 0.76
WS 0.25 0.79 0.08 0.32
-0.006
Multi-dimensional scaling (MDS) plot of pairwise FST values between
samples of Arripis truttaceus, as presented in Table 3.7. Sample codes are as per Table 2.1. The stress value is an inverse measure of the degree of correspondence between the distances among points implied by MDS and the matrix input.
75
Chapter 3. Population Genetic Structure Table 3.8
Results of tests for differentiation between pairs of samples of Arripis
georgianus presented as pairwise FST values (below diagonal) and the probability of obtaining by chance, a value as large or larger than the observed FST (above diagonal, italicised) using a permutation approach. Sample codes are as per Table 2.1. HF HF HC HP HI HO HW HS HF08* *
0.003 0.005 0.010 0.010 -0.016 0.004 0.002
HC 0.35 0.012 0.019 0.005 0.000 -0.002
HP 0.90 0.39 -0.013 -0.005 0.029 0.002
HI 0.34 0.14 0.92 -0.000 0.030 0.009
HO 0.37 0.38 0.68 0.30 0.026 -0.007
HW 0.87 0.61 0.24 0.07 0.19
HS 0.72 0.90 0.73 0.28 0.54 0.42
HF08* 0.69*
0.018
HF08 has only been compared to HF (see text).
Figure 3.3
Multi-dimensional scaling (MDS) plot of pairwise FST values between
samples of Arripis georgianus, as presented in Table 3.8. Sample codes are as per Table 2.1. The stress value is an inverse measure of the degree of correspondence between the distances among points implied by MDS and the matrix input. HF is the sample collected from Fremantle at the same time as the other samples (i.e. 2006) while HF08 is the sample collected from Fremantle in 2008. 76
Chapter 3. Population Genetic Structure
Table 3.9
Results of tests of differentiation in each Arripis species. (i) Single-locus
FST values across all samples for each species and the probability of obtaining, by chance, a value as large as or larger than the observed FST from 10,000 permutations. A dash (-) indicates that a single allele was recorded in the species. (ii) Multi-locus FST values from permutation test and the number of times a randomly permuted FST ≥ actual FST. Sample sizes are as reported in Table 3.2.
A. trutta FST p-value
A. truttaceus FST p-value
A. georgianus FST p-value
(i) Single-locus AldoB4 CaM4 CK7 S72
-0.010 -0.006 -0.002
0.34 0.52 0.63
0.011 0.001
0.23 0.42
0.007 -0.004 -0.009 0.005
0.14 0.98 0.26 0.48
(ii) Permutation
-0.005
0.73
0.005
0.22
0.005
0.10
77
Figure 3.4
Factorial correspondence analysis (FCA) illustrating the lack of strong relationship between the multi-locus
genotypes and collection site of individuals of Arripis trutta. The total variation explained by each axis is given as a percentage. Sample codes are as per Table 2.1. For scaling purposes, a single outlier has been shifted and its correct position indicated by an arrow and coordinates. Note, many points overlap and are obscured.
Figure 3.5
Factorial correspondence analysis (FCA) illustrating the lack of strong relationship between the multi-locus
genotypes and collection site of individuals of Arripis truttaceus. The total variation explained by each axis is given as a percentage. Sample codes are as per Table 2.1. For scaling purposes, outliers have been shifted and their correct position indicated by an arrow and coordinates. Note, many points overlap and are obscured.
Figure 3.6
Factorial correspondence analysis (FCA) illustrating the lack of strong relationship between the multi-locus
genotypes and collection site of individuals of Arripis georgianus. The total variation explained by each axis is given as a percentage. Sample codes are as per Table 2.1. For scaling purposes, outliers have been shifted and their correct position indicated by an arrow and coordinates. Note, many points overlap and are obscured.
Chapter 3. Population Genetic Structure 3.4
Discussion
This research did not reveal any evidence of genetic structuring in any of A. trutta, A. truttaceus and A. georgianus in Australian waters. For one or more species, this could be due to: (i) inadequate geographic sampling; (ii) errors in the genetic data; (iii) the effects of selection; and/or (iv) the presence of genetic structuring that was too subtle to be apparent given the loci/sample size combinations. Alternatively, and most likely, it could reflect the presence of sufficient gene flow across the entire range of the species to prevent genetic structuring. Each of these alternatives is considered in turn below.
Point (i) – inadequate geographic sampling.
The results were based on samples
obtained from virtually either the entire distribution of the species in Australian waters (A. trutta) or the entire distribution of the Australian species (A. truttaceus, A. georgianus). Thus, it seems highly unlikely that any genetic differentiation in any of the species in Australian waters was missed, as could have happened if, for example, all individuals had been sampled from a single, more localised breeding unit within a subdivided species (see Page and Holmes 1998).
Point (ii) – errors in the genetic data. Inaccurate results due to errors in the production and/or scoring of the genetic data seem unlikely for the following reasons. Firstly, most methodological errors are likely to produce heterogeneity rather than homogeneity (Pompanon et al. 2005), although the allele binning could have masked some variation (see Chapter 2). Secondly, the patterns of variation at each locus in each sample for each species were virtually always in accordance with those expected under HWE conditions. Thus, there was no evidence to suggest that a range of PCR artefacts, such 81
Chapter 3. Population Genetic Structure as null alleles and/or large allele drop out, which typically generate homozygote excesses (Pompanon et al. 2005), have had a major impact on the data sets. Thirdly, the patterns of polymorphism and genetic variation were consistent across multiple loci, suggesting there were no locus-specific errors. Finally, none of the positive or negative controls revealed any evidence of contamination of samples with non-target DNA or of scoring inconsistencies (beyond the boundaries of the allele bins).
Point (iii) – selection. The absence of genetic subdivision in the arripids is unlikely to be reflecting the effects of selection because the patterns of genetic diversity, genetic variance (FST) and heterozygosity were similar across multiple loci. Instead, selection is expected to act on an individual locus rather than the whole genome (Hare 2001).
Point (iv) – subtle genetic structure. A reliance on few loci and/or markers with low levels of polymorphism will limit the power of statistical tests (Ryman et al. 2006), and therefore the ability to detect genetic structuring, if it exists. In this regard, it is relevant that the detected amount of polymorphism at the intron loci in each species of Arripis was greater than that of earlier allozyme studies. In addition, power analyses showed that the minimum power of the present data to detect heterogeneity would be sufficient to detect genetic heterogeneity equivalent to an FST value of 0.009 to 0.039, depending on the species (see Section 3.3.4). Realistically, the only way to increase the power of the test for the presence of genetic structure in Arripis is to increase the amount of genetic information generated for each species.
For example, power analysis
simulations suggest that, in order to increase power by an order of magnitude using the current genetic markers, sample sizes would need to be increased at least six- to sevenfold (i.e. up to 700 – 1300 individuals per species; data not shown). 82
Chapter 3. Population Genetic Structure The ability to generate large amounts of genetic data for species, like those of Arripis, whose genome structures are largely unknown is becoming easier and cheaper (e.g. Santana et al. 2009). However, given the apparent extreme mobility of these fish and the limitations of genetic data in producing accurate information about high gene-flow situations, combined with the apparently low levels of genome-wide polymorphism in the representatives of this family (see Chapter 6), other non-genetic techniques (such as otolith microchemistry and tagging) may be potentially more useful in identifying subgroups, if they exist (see Point v, below).
Point (v) – no genetic structure. All of the available evidence suggests that each of the three study species of Arripis is genetically homogeneous throughout its entire distribution (A. truttaceus and A. georgianus), or entire Australian distribution (A. trutta). Neither the EPIC data of the current study nor the allozyme data of previous studies (i.e. A. truttaceus, MacDonald 1980; A. georgianus, Ayvazian et al. 2004) have revealed evidence of the presence of genetic structure in any of the study species. This is consistent with the observations of extensive movements in each species based on tagging (e.g. Stanley 1986; 1988a; 1988b; 1988c; Ayvazian et al. 2004; Stewart et al. 2011) and otolith microchemistry (Ayvazian et al. 2004; Hughes et al. 2011). These rigorous studies strongly suggest that the species of Arripis are highly vagile and do not spend their entire life in a single water body.
The apparent absence of genetic structure combined with the mobility suggested by tagging and otolith microchemistry data raises the possibility that one or more of the Arripis species are actually panmictic over their entire distribution (Australian species) or entire Australian distribution (A. trutta). Indeed, genetic data alone cannot be used to 83
Chapter 3. Population Genetic Structure prove or disprove a hypothesis of panmixia because very limited amounts of gene flow can homogenise the genetic composition of what are effectively independent breeding units in a species (Shaklee and Currens 2003). The European Eel (Anguilla anguilla) provides an excellent example of the difficulties of establishing whether or not a marine species is panmictic or even genetically homogeneous over its range. This eel shows a life-history consistent with high gene flow in that all adults (which are normally dispersed along the coasts of Europe and northern Africa) migrate to form a single breeding unit (in the Sargasso Sea), followed by larval and post-spawning adult dispersal.
For decades, based on life-history information as described above, the
European Eel was considered as a classic example of a panmictic species, and early studies did not find any evidence of genetic structuring across this species entire range (e.g. DeLigny and Pantelouris 1973; Lintas et al. 1998). However, three subsequent studies reported evidence of genetic structuring, including isolation-by-distance (Daemen et al. 2001; Wirth and Bernatchez 2001; Maes and Volckaert 2002). More recently, the results of carefully controlled studies have reinstated the original idea that panmixia is the most likely model and suggested that the structure detected in some of the previous studies resulted from temporal variation in recruitment and sampling (Dannewitz et al. 2005; Palm et al. 2009; Als et al. 2011).
The results of this study add to a small list of pelagic coastal species that have been found to be genetically homogeneous and possibly panmictic over a vast distance (e.g. Hoolihan et al. 2006; Cárdenas et al. 2009; Palm et al. 2009). In the case of each Arripis species, this high gene flow life-history is closely coincident (and possibly linked) with oceanographic conditions where larval dispersal is passively driven by
84
Chapter 3. Population Genetic Structure winds and currents, followed by active movements of immature and adult fish that seasonally converge on a spatially restricted breeding ground (see Introduction).
Finally, the distribution of A. trutta extends beyond Australian waters into New Zealand. Despite repeated attempts to obtain samples from New Zealand, only two specimens from near Christchurch (43° 32' S, 172° 37' E) could be sourced. The alleles present in these two specimens (EZ01: 178/191, 919/919, 961/961; EZ02: 178/178, 914/914, 943/952 for loci AldoB4, CK7 and S72, respectively) were among the common alleles in the samples from eastern Australia. While this might suggest the presence of recent (in an evolutionary sense) connections between the A. trutta in Australia and New Zealand, obviously the sample size is inadequate to draw any conclusions about the extent of any mixing. However, such a connection might be expected based on water movement of the East Australian Current, which delivers a branch of water (and thus possibly larvae) from the spawning area of A. trutta in New South Wales directly to the west coast of New Zealand (Bostock et al. 2006; see Chapter 5).
3.4.1
Conclusions
This study provides the strongest evidence to date that each of Arripis trutta, A. truttaceus and A. georgianus is genetically homogenous in Australian waters, although the results are to a certain extent limited by modest sample sizes and a small number of polymorphic loci. In fact, the combination of genetic and life-history data suggest that each species of Arripis has considerable gene flow throughout these waters and might, in fact, comprise a single, panmictic population. On the basis of these results, it is valid to treat each species as a single population in analyses in subsequent chapters. 85
Chapter 4. Molecular Taxonomy
4.0
A Molecular Approach to the Taxonomy of Australian Salmon
4.1
Introduction
As discussed in Chapter 1, one of the main objectives of this thesis was to use molecular data to investigate the evolutionary relationships of the putative geminate species of A. trutta and A. truttaceus. However, it is relevant to first establish whether genetic data are consistent with the current view in the literature (e.g. Paulin 1993; Gomon et al. 2008), that these two taxa represent ‘true’ species (as opposed to conspecific populations that are exchanging genes or have recently done so). Since the taxonomic status of these two taxa is linked with that of the third species of Australian Salmon, A. xylabion, the status of this latter taxon is also a critical part of this issue.
The current view that each of the three types of Australian Salmon – A. trutta, A. truttaceus and A. xylabion – are separate species is based on morphological evidence. In particular, A. truttaceus differs from each of the other species in having fewer lower gill rakers and from A. xylabion (but not A. trutta) by having fewer upper gill rakers and a longer upper lobe of the caudal fin (Table 4.1; Figure 4.1). The length of the upper lobe of the caudal fin is also longer in A. xylabion than in A. trutta (Table 4.1; Figure 4.1). There are no other (known) diagnostic differences between A. trutta and A. xylabion, or between A. trutta and A. truttaceus (Table 4.1; Figure 4.1). Note the Australian Herring A. georgianus is distinguished from the three Australian Salmon by its much smaller maximum length and higher gill raker count (Table 4.1). As implied 86
Chapter 4. Molecular Taxonomy above, the extent of the morphological differences between the Australian Salmon is limited and subtle. Consequently, the taxonomic status of the three species has changed through time, where various workers have recognised the Australian Salmon as a single species, two to three subspecies, or, two to three full species (see Paulin 1993).
Table 4.1
Selected morphological and meristic measures of n individuals of the
four species of Arripis. Data from Paulin (1993). Also see Figure 4.1. A. trutta
A. truttaceus
A. xylabion
A. georgianus
34
20
32
7
Maximum length
890 mm
960 mm
850 mm
410 mm
Gill raker count†
9-13 + 20-24
7-11 + 16-17
12-17 + 20-25
16-18 + 28-32
Dorsal fin rays#
IX, 15-16
IX, 15-16
IX, 15-16
IX, 13-14
Anal fin rays#
III, 9-10
III, 9-10
III, 9-10
III, 10
Pectoral fin rays
16-18
16-18
16-18
16-18
Lateral line scales
49-53
49-56
50-55
54-59
25
25
25
25
28.2% (22.8 – 29.7)
26.2% (23.7 – 28.3)
34.1% (30.4 – 37.7)
29.3% (26.1 - 31.9)
n*
Vertebrae§ Caudal lobe^ *
number of specimens examined by Paulin (1993) gill raker counts of first gill arch given as a range in the format: upper + lower (including rudiments) # fin ray counts given as spines (Roman numerals), soft rays (Arabic numerals) § total number of vertebrae ^ mean length of upper lobe of caudal fin as percentage of standard length (range in parentheses) †
Modern approaches to species delimitation typically include genetic evidence (along with morphological evidence) because morphological data are not always sufficient for species recognition.
For example, some species are polytypic with considerable
morphological variation (e.g. Hull et al. 2010), while, on the other hand, cryptic species may be morphologically indistinguishable (e.g. Ross et al. 2010). Having said this, it is 87
Chapter 4. Molecular Taxonomy
Figure 4.1
Graphs showing the range of measurements for the three diagnostic
characters in Australian Salmon – Arripis trutta (black), A. truttaceus (grey), A. xylabion (stippled). Bars show the range of measurements in each species for: A. number of upper gill rakers; B. number of lower gill rakers; and C. the length of the upper caudal lobe as a percentage of standard length (mean shown as vertical bar). Data from Paulin (1993). See also Table 4.1. 88
Chapter 4. Molecular Taxonomy also important to recognise that genetic data do not invariably allow for the unambiguous resolution of species boundaries due to a range of problems, such as incomplete lineage sorting (e.g. Avise and Ball 1990; Knowles and Carstens 2007), hybridisation and introgression (e.g. Mallet and Willmott 2003) and homoplasy (e.g. Xia 2009). In addition, a gene tree might not reflect the species tree, especially when only a single gene is used (e.g. Mallet and Willmott 2003).
Genetic (and morphological) approaches to species delimitation are to a certain extent plagued by the controversies surrounding the appropriate conceptual definition of a species and the operational criteria for recognising the species (Sites and Marshall 2004). Rather than overtly committing to a particular species concept, many recent studies in molecular taxonomy, including the present chapter, have adopted part of the approach of de Queiroz (1998) and others (e.g. Mayden 1999; Hey 2006a), whereby any of the properties acquired by lineages during divergence (e.g. reciprocal monophyly, fixed allelic differences) are potentially informative and multiple pieces of supportive evidence are required.
Monophyly, or some other form of genetic exclusivity (where all members of a species form a group, to the exclusion of all members of other species), is probably the most widely used genetic criterion for species delimitation (reviewed in Sites and Marshall 2004). This criterion is broadly acceptable because lineages are expected to reach this condition only after a relatively long period of isolation (lineage sorting) (see Avise and Ball 1990; Knowles and Carstens 2007). In this sense, genetic exclusivity provides a relatively conservative test of species boundaries. In addition, it is applicable to a broad range of taxa (Knowles and Carstens 2007). Monophyly is often tested using mtDNA 89
Chapter 4. Molecular Taxonomy gene trees, partly because a mtDNA locus will generally reach monophyly faster than a nDNA locus due to it smaller effective population size (Avise and Ball 1990). However, phylogenetic reconstruction based on a single locus is potentially misleading and there is consensus that genetic methods of species delimitation should incorporate loci from both mitochondrial and nuclear genomes (e.g. Sites and Marshall 2004; de Queiroz 2007; Knowles and Carstens 2007). Other types of genetic properties that are commonly used to distinguish species include fixed allelic differences or differences in allele frequencies (e.g. Paine et al. 2007) and/or the presence of interspecific differences that are substantially greater than intraspecific differences (e.g. Ward 2009).
The only previous attempt to use genetic data to address taxonomic questions within the Arripidae is in the PhD research of MacDonald (1980). This research used allozyme markers to address, inter alia, the species status of the east and west coast populations of Australian Salmon. The results supported the view that these populations represent different species because fixed allelic differences were found at two of 27 loci. However, MacDonald’s work: (i) is unpublished; (ii) did not include specimens from the ‘population’ to which the name A. xylabion is now applied (it pre-dated Paulin’s (1993) revision of the family); and (iii) was restricted to allozyme markers. Hence, there is a need for an analysis of the taxonomic status of the group that includes all three species of Australian Salmon and uses additional genetic markers.
4.1.1
Aim
The present chapter investigates whether the current, morphologically-based taxonomy of Australian Salmon (described above) is supported by DNA data. The molecular data 90
Chapter 4. Molecular Taxonomy will be considered to support the current taxonomy of Arripis if there is reciprocal monophyly of species haplotypes in the mtDNA trees and if this monophyly is accompanied by other lines of supportive evidence (e.g. fixed allelic differences and/or significant allele frequency differences), especially in the nuclear genome. In addition, the interspecific genetic distances are expected to be substantially greater than intraspecific ones.
4.2
Materials and Methods
4.2.1
Sampling Regime
The results of this chapter are based upon the analysis of samples of: A. trutta from four sampling sites (New South Wales, Victoria, Tasmania, New Zealand); A. truttaceus from five sampling sites ranging from Augusta, Western Australia to The Coorong, South Australia; A. xylabion from Lord Howe Island; and the outgroup species A. georgianus from six sampling sites ranging from Fremantle, Western Australia to The Coorong, South Australia (see Chapter 2 for details). The sample sizes for each of the sampled mtDNA and intron loci for each sampling site for each species are provided in Table 4.2. Given no evidence of population structure in any species (Chapter 3), individuals from different sites were pooled into a single sample for each species for the phylogenetic analyses. This included two individuals of A. trutta from New Zealand, which were pooled with the Australian samples of this species on the basis that the haplotypes and alleles in these fish were among the most common in A. trutta from Australia (Chapter 3, Section 3.4). Details of the sampling sites and the collection methods were provided in Chapter 2. 91
Table 4.2
Sample sizes for each gene used in the phylogenetic analyses of Arripis. The total (pooled) number of individuals used in
the analyses is provided (bold), along with the sample sizes for each sampling site for each species of Australian Salmon (ingroup). Sample site abbreviations are the same as those used throughout the thesis and are defined in Chapter 2. To simplify presentation, only total sample sizes are given for the outgroup taxon A. georgianus (details of geographic sampling strategy are given in Chapter 2).
Locus
Arripis trutta EN EV ET EZ
Mitochondrial Loci cytochrome b 8 8 COI Nuclear Loci 29 AldoB4 CaM4 * 3 28 CK7 29 S72
Total
Arripis truttaceus WA WY WL WW WS Total
Arripis xylabion KL
Arripis georgianus (outgroup)
7 7
8 8
2 2
25 25
7 7
7 7
6 6
8 8
8 8
36 36
11 11
1 1
30 3 28 25
30 3 29 28
2 2 2 2
91 11 87 84
30 2 8 27
15 2 9 15
39 2 8 38
9 2 12 13
37 2 10 45
130 10 47 138
25 8 20 21
226 222 211 195
* Data for CaM4 are included here, foreshadowing their use in analyses in Chapter 5.
Chapter 4. Molecular Taxonomy 4.2.2
Genetic Assays
The mitochondrial DNA data used in this chapter were in the form of the nucleotide sequence of a 483 bp portion of the cytochrome b gene and a 654 bp portion of the cytochrome oxidase sub-unit 1 gene (COI) in Arripis. These data were generated using the methods described in Chapter 2. For each gene, all sequences were the same length (i.e. with no insertions or deletions) in all sampled individuals and there were no missing data (i.e. unresolved nucleotides). The nuclear DNA data used in this chapter were in the form of the frequency of length variants (alleles) in the introns of three genes, selected on the basis that they were polymorphic in at least one of the three species of Australian Salmon (see Table 3.2). These nDNA data were generated using the methods described in Chapter 2 and were also used to assess the population structures of each of these species in Australian waters, as described in Chapter 3.
4.2.3
Data Analyses
4.2.3.1
mtDNA Sequence Characteristics
The general characteristics of the sequences of each mitochondrial region were assessed using MEGA version 4 (Tamura et al. 2007). Firstly, the average composition of the nucleotide sequence was calculated as a percentage of each base (A, T, C, G). Secondly, for each gene, the number of polymorphic sites along the aligned sequences was counted, along with the number of parsimony informative sites.
Finally, the
transition/transversion ratio was determined.
93
Chapter 4. Molecular Taxonomy 4.2.3.2
Monophyly of mtDNA Sequences using Phylogenetic Analyses
Phylogenetic approaches were used to assess monophyly of the mtDNA sequences of each species. Before providing details on these analyses, it is first necessary to address four issues that underlie the phylogenetic analyses.
4.2.3.2.1
Choice of outgroup
Arripis georgianus was chosen as the outgroup taxon to provide a reference point for resolving the relationships among the Australian Salmon (see Maddison et al. 1984). It was chosen because it is the only other extant member of the family (Paulin 1993), and diverged prior to divergences within the ingroup (MacDonald 1980). Specifically, concatenated sequence data (see Section 4.2.3.3.2) from a single individual A. georgianus (HC01, from Canal Rocks, Western Australia) were included in the mtDNA-based analyses, while the allele frequency data at the intron loci were from the individuals of A. georgianus assayed as a part of the population genetic structure analysis in Chapter 3.
4.2.3.2.2
Concatenated sequence data
The cytochrome b and COI sequences of each individual were concatenated. Concatenation was used here because the concatenated sequences were expected to provide a better phylogenetic signal by increasing the number of segregating sites (e.g. Cunningham 1997a; 1997b; Ballard et al. 1998; Gadagkar et al. 2005). To assess the validity of combining the data for the two regions of the mtDNA a partition94
Chapter 4. Molecular Taxonomy homogeneity test (or incongruence length difference test; Farris et al. 1994; 1995), was implemented in PAUP* using a branch and bound search over 1000 replicates. This test evaluated whether the two gene regions contained similar phylogenetic signals by measuring whether the likelihood (of a tree) increased when loci were combined, compared to random data partitions. Since there was no incongruence in the signal between the cytochrome b and COI data (p = 1.00), the sequences were concatenated without partitions (total 1137 bp) for phylogenetic analyses. Throughout this chapter, the concatenated mtDNA sequence of an individual is referred to as its haplotype.
4.2.3.2.3
Substitution saturation
Tests for substitution saturation in the concatenated mtDNA sequence data were conducted among all haplotypes in all species because multiple substitutions at a site, especially backward substitutions may create a ‘saturated’ site and determining the relatedness between sequences becomes problematic due to homoplasy (Nei and Kumar 2000; Xia et al. 2003). Using the entropy method in the software DAMBE (Xia et al. 2003), an index (Iss) was calculated whereby the mean entropy (uncertainty due to randomness) among sites was compared against the expected entropy under full sequence saturation (Xia 2009). A two-tailed t-test was used to determine whether the Iss was significantly smaller than the critical index value (Iss.c) at which sequences began to fail to recover the tree (Xia 2009). The results did not reveal any evidence of saturation among concatenated mtDNA sequences of the four species of Arripis (Iss = 0.05 < Iss.c = 0.77; p = 0.00).
95
Chapter 4. Molecular Taxonomy 4.2.3.2.4
Substitution model
The substitution model of Tamura & Nei (1993) with a gamma distribution correction, G (Uzzell and Corbin 1971), was determined to be the most appropriate for use in the phylogenetic analyses with the concatenated mtDNA sequence data. This model (TrN + G) was selected using concatenated mtDNA sequence data for all haplotypes for all species using the Akaike Information Criterion (AIC; Table 4.3) and the Bayesian Information Criterion (BIC; results not shown), as implemented in JMODELTEST version 0.1.1 (Guindon and Gascuel 2003; Posada 2008). Tamura & Nei’s (1993) substitution model allows for different transversion and transition rates, and also makes
Table 4.3
Output of JMODELTEST showing the values of relevant parameters for the
Tamura-Nei (with gamma correction) substitution model based on Aikake Information Criterion (AIC) on the concatenated mtDNA data set using all haplotypes for all species of Arripis. The frequency of each of the bases (Freq), values for the rate matrix (R) and the gamma shape correction value (G) are provided.
Parameter Model Selected Freq [A] Freq [C] Freq [G] Freq [T] R [AC] R [AG] R [AT] R [CG] R [CT] R [GT] G
Value TrN + G 0.2441 0.3024 0.1616 0.2920 2.4791 24.0650 1.4160 0.2171 14.7458 1.0000 0.2718 96
Chapter 4. Molecular Taxonomy a distinction for transition rates between purines and between pyrimidines, while the gamma distribution corrections account for substitution rate variation among codon positions (Uzzell and Corbin 1971). In phylogenetic analysis, it is important to apply an appropriate substitution model because otherwise various changes within a sequence, such as backward or parallel substitutions and unequal rates of substitution, may complicate the interpretation of homology (Nei and Kumar 2000).
4.2.3.2.5
Phylogenetic analyses
Having established the preliminary parameters and assumptions above, the phylogenetic relationships between the three species of Australian Salmon, based on the 1137 bp concatenated mtDNA sequences were examined by four methods: (i) distance (phenetic); (ii) maximum parsimony; (iii) maximum likelihood; and (iv) Bayesian inference (Felsenstein 1988; Nei and Kumar 2000).
Although used to assess
monophyly in the present chapter, all methods ultimately represent phylogenetic relationships in a tree (and are used for such in Chapter 5). Branch tips of the tree represent taxa and the patterns of branching within the tree represent relationships between the taxa (Swofford et al. 1996). All methods were included because each has strengths and weaknesses, as discussed below.
4.2.3.2.5.1 Distance-based method
A matrix of pairwise TrN + G distances between haplotypes of all species of Australian Salmon was generated in PAUP*. From this matrix, a 50% majority rule consensus neighbour-joining tree, based on 1000 bootstrap replicates, was constructed with A. 97
Chapter 4. Molecular Taxonomy georgianus as the outgroup. Although phenetic/distance-based measures have largely been superseded (Felsenstein 1988; Rohlf et al. 1990; Li and Gouy 1991; Nei 1991; Hillis et al. 1994), a neighbour-joining analysis was included in this study because it is believed to produce reliable results when sequences are similar (which is the case for the Arripis data) and because it is based on very few assumptions (Saitou and Nei 1987; Swofford et al. 1996; Nei and Kumar 2000).
Support for neighbour-joining trees was assessed by bootstrapping, which provided the frequency (or, bootstrap value) with which a particular node appeared in 1000 replicates using a subset of characters (nucleotides) from the original data matrix (Felsenstein 1985). Thus, this method resampled characters rather than taxa (Felsenstein 1988). Since bootstrapping generated numerous resampled trees, a majority rules consensus tree was constructed to include nodes that were present in ≥ 50% of all generated trees (Felsenstein 1988).
4.2.3.2.5.2 Maximum parsimony
Maximum parsimony analyses were conducted using unweighted parsimony as implemented in the software PAUP* with the three Australian Salmon as the ingroup and A. georgianus as the outgroup. A maximum parsimony consensus tree, based on 1000 bootstrap replicates, was generated by a heuristic search over 1000 replicates and tree bisection and reconnection branch swapping (see below).
Although maximum parsimony analysis makes use of informative characters only; in this case, derived nucleotide substitutions that were shared among a particular subset of 98
Chapter 4. Molecular Taxonomy Arripis haplotypes (i.e. synapomorphies; Hennig 1965; Fitch 1977; Swofford et al. 1996), it was included in this study because it is free from many of the assumptions of other methods such as a particular substitution model (Nei and Kumar 2000) (although not entirely assumption-free; e.g. branch lengths are assumed to be equal, Lewis 2001). Also, when sequence divergence is low (as in the Arripis data), the rate of nucleotide substitution is approximately constant and, when a large number of nucleotides is examined, maximum parsimony methods often perform better than distance methods (Sourdis and Nei 1988; Nei 1991).
In this analysis, the length of each possible tree topology was calculated as the minimum number of substitutions over all sites, and the topology with the shortest tree length was considered the best, or most parsimonious tree (Nei and Kumar 2000). The tree bisection and reconnection (TBR) method was used to generate possible tree topologies because it is the most extensive method available in PAUP* (Swofford and Sullivan 2009) and among the most commonly used true heuristic searches.
4.2.3.2.5.3 Maximum likelihood
Maximum likelihood analyses (Felsenstein 1973; 1981), were conducted using the TrN + G model, as implemented in the software PAUP*, with the three Australian Salmon as the ingroup and A. georgianus as the outgroup. Values for base frequencies, rate matrix and gamma shape values were specified from the JMODELTEST output (Table 4.3). A maximum likelihood consensus tree, based on 1000 bootstrap replicates, was generated by a heuristic search and TBR branch swapping.
99
Chapter 4. Molecular Taxonomy Maximum likelihood was used in this study because it is a consistent estimator of tree topologies with a lower variance than other methods (Felsenstein 1973; Swofford et al. 1996; Nei and Kumar 2000). This method also utilised the whole sequence, including singletons, rather than just ‘informative’ sites (as in maximum parsimony). In addition, the maximum likelihood analysis explicitly allowed for heterogeneity in branch lengths whereas that based on maximum parsimony implicitly assumed equal branch lengths (Lewis 2001).
Maximum likelihood analysis was used to find the tree topology with the highest likelihood of observing the given data (Nei and Kumar 2000). Starting with a random unrooted tree topology (based on the data), the tree topology was changed incrementally and the likelihood (average likelihood of each nucleotide arising independently at each position in the sequence for each taxon; Felsenstein 1981) for the new tree calculated. When the likelihood of the new tree was greater, the change was accepted and the new tree became the basis for further changes (Felsenstein 1973).
4.2.3.2.5.4 Bayesian inference
A 50% majority-rule consensus tree was constructed using Bayesian inference, as implemented in the software MRBAYES version 3.1.2 (Ronquist et al. 2009), with the three Australian Salmon as the ingroup and A. georgianus as the outgroup. Using the general time reversible model with gamma corrections (GTR + G; Rodriguez et al. 1990), default Markov Chain Monte Carlo (MCMC) settings (one cold and two heated chains, heating temperature = 0.2) were run for one million generations with a sampling frequency of 100 generations. The first 25% of samples and trees were discarded as 100
Chapter 4. Molecular Taxonomy burn-in as recommended (Ronquist et al. 2005). Support for the consensus tree was summarised by the proportion of credible trees exhibiting the most common node splits (i.e. posterior probabilities; Ronquist et al. 2009). The MCMC process (see below) visits trees in proportion to their posterior probability, and so this value is a natural measure of nodal support, i.e. a true probability tree (Lewis 2001; Ronquist et al. 2009).
Bayesian inference was included in addition to the maximum likelihood method because it approaches the question from a different direction - by determining the probability of the hypothesis given the data, rather than the probability for the data given the hypothesis (Nei and Kumar 2000; Lewis 2001). Thus, it is the probability of x occurring, conditioned on y (Huelsenbeck et al. 2002; Ronquist et al. 2009), where the posterior probability of a tree, x, is related to the likelihood of the observed data which is determined by the prior probability of the tree, y (such as the GTR + G model of substitution).
The posterior probability is proportional to the product of the prior probability and the likelihood (Ronquist et al. 2009). The strength of the estimate of prior probability is largely dependent on the substitution model (Lewis 2001; Ronquist et al. 2009). The GTR + G model, rather than the TrN + G model, was used in this analysis because: i) the AIC scores for the two models were almost identical; and ii) MRBAYES does not incorporate the TrN model, and the GTR is the most parameterised alternative it offers and is the recommended default model (Tavare 1986; Ronquist et al. 2005).
Unlike the previous methods, a single optimal tree is not possible in Bayesian analyses. Instead it used a MCMC simulation to sample trees in proportion to their likelihood and 101
Chapter 4. Molecular Taxonomy generate a sample of credible trees that was used to create a consensus tree (Huelsenbeck et al. 2002). In a chain-like fashion, MCMC searched for peaks on the likelihood surface, where each point on the surface represented a likelihood value for a given hypothetical tree.
By randomly moving across the likelihood surface and
accepting the move if likelihood improved, chains ascended likelihood peaks to identify localised optima (Ronquist et al. 2009). To cover a greater area of the likelihood surface, the Metropolis coupled-MCMC algorithm was used, with several chains simultaneously searching and randomly switching positions (Ronquist et al. 2005).
4.2.3.2.6
Parsimony networks
A parsimony (haplotype) network was constructed with default settings (95% confidence) in the software TCS version 1.21 (Clement et al. 2000) using the concatenated mtDNA sequence data. This method evaluated the minimum number of substitutions required to connect parsimoniously any two haplotypes. A network was constructed by firstly linking sequences with the smallest number of differences and then connecting to sequences with increasing numbers of differences. The original concept of parsimony networks (Templeton et al. 1992; 1995) was rooted in intraspecific phylogeographic inference, particularly nested-clade analysis (NCA; Templeton 1998), much of which is now out of favour (e.g. Knowles 2008; Petit 2008). Despite this, parsimony networks provide a useful method for displaying relationships between haplotypes of a single species and also for interpreting interspecific relationships when the taxa are closely related (Templeton 2004).
102
Chapter 4. Molecular Taxonomy 4.2.3.3
Fixed Differences Between mtDNA Sequences
Fixed nucleotide sites among aligned concatenated mtDNA sequences were identified using MEGA. In this chapter, a fixed site is defined as a site with a nucleotide common to all sequences in one species, but with a different nucleotide(s) at the site in the other species of Australian Salmon involved in the comparison (Nei and Kumar 2000).
4.2.3.4
Genetic distances based on mtDNA sequences
A matrix of pairwise DA distances (Nei et al. 1983) based on TrN + G substitution model between the different mtDNA haplotypes was generated in PAUP* version 4.0b10 (Swofford 2005). The corrected chord distance DA was used because it is not biased by low-frequency haplotypes and low sample size (Nei et al. 1983).
4.2.3.5
Fixed Differences in nDNA Genotypes
A table of nDNA allele frequencies (Table 3.2) was assessed for fixed allele differences, defined as an allele present at a locus in one species, but not found at that locus in either of the other species of Australian Salmon (Nei and Kumar 2000).
4.2.3.6
Differentiation in nDNA Allele Frequencies
Exact probability tests of differentiation (G-tests) were used to test for the presence of statistically significant allele frequency differentiation at each intron locus between each pair of species of Australian Salmon. These tests used a Markov chain, as implemented 103
Chapter 4. Molecular Taxonomy in GENEPOP (Raymond and Rousset 1995a), to estimate the exact probability of a Type I error. This type of test was used because it is not biased by rare alleles (Raymond and Rousset 1995b), which were widespread in the Arripis data (see Results). In addition, using GENEPOP, a Fisher’s exact test was used to test for evidence of differentiation across loci for each pairwise combination of the three species of Australian Salmon.
4.2.3.7
Factorial Correspondence Analysis using Multilocus nDNA Genotypes
A factorial correspondence analysis (FCA) was performed on the multilocus genotypes of individuals using the three intron loci that were polymorphic in at least one of the species of Australian Salmon (AldoB4, CK7, S72).
The FCA was conducted in
GENETIX, as described in Section 3.2.3.6.
4.3
Results
4.3.1
Mitochondrial DNA
The average base composition of the mtDNA sequences was similar for all species of Australian Salmon (Table 4.4). It was slightly biased towards adenine and thymine, as has been found for the mtDNA in other teleosts (Saccone et al. 1999; Ward et al. 2005). Alignment of the cytochrome b sequences from the different individuals of the three species of Arripis revealed 16 haplotypes and a relatively low level of variability, with 17 of the total of 483 sites variable among species (Tables 4.4, 4.5). Of the 17 variable
104
Table 4.4
Selected characteristics of the 483 bp partial cytochrome b sequences and 654 bp partial COI sequences in three species of Australian
Salmon. For each gene region, the number of individuals assayed, the number of haplotypes, the average base composition of the sequence (expressed as percentage), the number of polymorphic sites (i.e. where nucleotide differs between sequences), the transition /transversion (Ts / Tv) ratio, and the codon positions at which the polymorphic sites are found, are presented. Data incorporating sequences from all species are indicated in bold. See Table 4.5 for full details of polymorphic sites.
cytochrome b A. trutta A. truttaceus A. xylabion Total
%G
Number of polymorphic sites
Ts / Tv ratio
Codon position (1 / 2 / 3)
25.3 25.5 25.7 25.4
13.7 13.5 13.3 13.5
11 3 0 17
10 / 1 3/0 16 / 1
2/1/8 0/0/3 2 / 1 / 14
24.9 24.9 25.0 24.9
16.8 16.8 16.8 16.8
8 1 1 15
7/1 1/0 1/0 14 / 1
0/0/8 0/0/1 0/0/1 0 / 0 / 14
Number sequenced
Number of haplotypes
%T
%C
%A
25 36 11 72
11 4 1 16
28.1 28.0 28.0 28.0
32.9 33.1 33.1 33.1
25 36 11 72
8 2 2 11
31.1 30.9 30.7 30.9
27.2 27.4 27.5 27.3
COI A. trutta A. truttaceus A. xylabion Total
Chapter 4. Molecular Taxonomy sites, almost all (i.e. 16/17) showed transitional changes and most were at third codon positions (Tables 4.4, 4.5), as would be expected for comparisons between closely related species (Nei and Kumar 2000). Likewise, the cytochrome oxidase sub-unit 1 (COI) gene, which was sequenced for all of the same individuals, revealed a total of 11 haplotypes (Tables 4.4, 4.5). The level of variability in these sequences was lower than that of cytochrome b, with 15 of the total of 654 sites variable among species (Tables 4.4, 4.5). All of the polymorphic sites showed transitional changes at third codon positions (Tables 4.4, 4.5). The concatenated mtDNA sequence was 1137 base pairs long and resulted in 14 A. trutta haplotypes, five A. truttaceus haplotypes, and two A. xylabion haplotypes (Table 4.5). Alignment of the 72 concatenated mtDNA sequences from the different individuals of the three species of Arripis revealed 32 polymorphic sites among species (Table 4.5).
Table 4.5
(over page) The location and distribution of 32 polymorphic sites in the
concatenated mtDNA sequences (483 bp portion of the cytochrome b gene and 654 bp portion of the COI gene) in three species of Australian Salmon. These polymorphic sites revealed a total of 21 haplotypes: 14 in A. trutta; 5 in A. truttaceus; and 2 in A. xylabion (Roman numerals). A dot indicates that the nucleotide at that position in a haplotype is the same as that indicated for Haplotype I. Fixed differences for each species, when present, are indicated in bold. Sample sizes (n) and the frequency of each haplotype (f) for each species are also provided. See Table 4.4 for a summary of selected characteristics of the sequences. 106
174
180
198
204
206
228
276
286
291
306
307
345
357
378
468
39
180
270
273
324
339
360
468
486
489
504
507
564
597
645
Codon position
141
f
COI
87
Haplotype Number
cytochrome b
3
3
3
3
3
3
2
3
3
1
3
3
1
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
T · · · · · · · · · · · · ·
A · · · · · · · · · · · · ·
T · · · · · · · · · · · · ·
G · · · · · · · · · · · · A
C · · · · · · · · · · · · ·
A · · · · · · · · · · · G ·
T · · · · · · · · · C C ·
C · · · · · · · · A · · · ·
T · · · · · · · · · · · · ·
C · · · · · · · · · · T · T
A · · · · · · · G · · · · ·
T · · · · · · C · · · · · ·
A · · · · · · G · · · · · ·
G · · · · · A · · · · · · ·
A · · · · G · · · · · · · ·
G · · · A · · · · · · · · ·
C · · · T · · · · · · · · ·
A · · · · · · · · · · · · ·
C · T T · · · · · · · · · ·
A · · · · · · · · · · · · ·
C · · · · · · · · · · · · ·
C · A A · · · · · · · · · ·
G A · · · · · · · · · · · ·
T · · · · · · · · · · · · ·
T · · · · · · · · · · · · ·
G · · · · · · A · · · · · ·
A · · · · · G · · · · · · ·
G · · · · · · · · · · · · ·
G · · A · · · · · · · · · ·
T · · · · C · · · · · · · ·
A · · · · · · · · · · · · ·
C C C C C
· · · G ·
C C C C C
· · · · ·
T T T T T
· · · · ·
. . C · ·
· · · · ·
C C C C C
T T T T T
· · · · ·
· · · · ·
· · · · ·
· · · · ·
· · · · ·
· · · · ·
· · · · ·
· · · · ·
· · · · ·
· G · · ·
T T T T T
· · · · ·
· · · · ·
C C C C C
C C C C C
· · · · ·
· · · · ·
A A A A A
A A A A A
· · · · ·
G G G G G
C C
· ·
· ·
· ·
· ·
· ·
· ·
· ·
C C
T T
· ·
· ·
· ·
A A
· ·
· ·
· ·
G ·
· ·
· ·
· ·
· ·
· ·
C C
C C
· ·
G G
A A
A A
· ·
· ·
A. trutta (n = 25) I II III IV V VI VII VIII IX X XI XII XIII XIV
0.48 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04 0.04
C · · · · · · · · · · · · ·
A. truttaceus (n = 36) XV XVI XVII XVIII XIX
0.89 0.03 0.03 0.03 0.03
· · · · T
A. xylabion (n = 11) XX XXI
0.82 0.18
· ·
Chapter 4. Molecular Taxonomy 4.3.1.1
Monophyly of mtDNA Sequences using Phylogenetic Analyses
Three of the four phylogenetic tree building methods using concatenated mtDNA sequences produced monophyletic terminal clusters. Each cluster included only the haplotypes of a single species, with nodal support above 70% in each case (Figure 4.2). The exception was maximum likelihood analysis in which a monophyletic clade for the two A. xylabion haplotypes was present in only 35% of bootstrap replicates (and therefore not represented in the 50% majority consensus tree in Figure 4.2C). In addition, the parsimony network showed complete monophyly of the concatenated mtDNA haplotypes of each species (Figure 4.3), with at least five mutational steps between the haplotypes of the different species (Figure 4.3). Similar results were found with each method when each gene was analysed separately (full data not shown).
Figure 4.2
(over page) Majority consensus phylogenies of Australian Salmon
inferred from concatenated mtDNA sequence haplotypes based on four methods. A. Neighbour joining (NJ) method using distances based on the TrN + G model. B. Maximum parsimony (MP; bootstrap values above line), utilising a heuristic search with TBR branch swapping. Bayesian inference (BI), using the GTR + G model generated the same tree (posterior probabilities below line). C. Maximum likelihood (ML), using the TrN + G model with a heuristic search and TBR branch swapping. Clades are colour coded by species. Haplotypes are numbered as in Table 4.5. Node support values were derived from 1000 bootstrap replicates (NJ, MP, ML) or 15,000 sampled post burn-in trees (BI). In each tree, the congener Arripis georgianus was the outgroup. 108
Chapter 4. Molecular Taxonomy
Figure 4.3
(over page) Parsimony network for the mtDNA haplotypes from the
concatenated sequence data (cytochrome b and COI) of the three species of Australian Salmon obtained by the criterion of parsimony with the program TCS version 1.21 (Clement et al. 2000). Each line in the network shows a single mutational change. A haplotype is shown by a coloured circle, the surface area of which is proportional to the frequency of the haplotype. Intermediate, unsampled haplotypes are indicated by a small black dot. Similar networks are shown for each of the genes separately (inset). In each case, Arripis trutta is represented by yellow; A. truttaceus by pink; and A. xylabion by orange. 110
Chapter 4. Molecular Taxonomy 4.3.1.2
Fixed Differences Between mtDNA Sequences
The concatenated mtDNA sequences of all three species of Australian Salmon showed fixed differences at multiple sites, in pairwise comparisons with each of the other species (Tables 4.5, 4.6).
In addition, five sites in A. trutta and four sites in A.
truttaceus were monomorphic within the species but had a nucleotide that was not present at that site in the sequences any other species (Tables 4.5, 4.6).
Table 4.6
Number of fixed site differences in the 1137 bp concatenated mtDNA
sequences of each species of Australian Salmon. A fixed site had a nucleotide common to all sequences in a species, but a different nucleotide(s) in the other species involved in the comparison. Data are presented as fixed differences in each species (bold) relative to the other two species combined and fixed differences in pairwise comparisons between two species.
A. trutta A. truttaceus A. xylabion
4.3.1.3
Fixed Differences
A. trutta
A. truttaceus
5 4 0
9 5
6
Intraspecific versus Interspecific Genetic Distances in mtDNA
Using concatenated mtDNA sequences, all pairwise interspecific distances, DA, were low in absolute terms, but were nevertheless between 4 and 12 times greater than the respective intraspecific DA genetic distances (Table 4.7).
112
Chapter 4. Molecular Taxonomy Table 4.7
Average pairwise DA genetic distances (and standard error) among n
different mtDNA haplotypes within each Arripis species (in italics along the diagonal) and between each species, expressed as a percentage.
A. trutta (n = 14) A. truttaceus (n = 5) A. xylabion (n = 2)
4.3.2
A. trutta 0.29 ± 0.01 1.30 ± 0.01 1.06 ± 0.02
A. truttaceus
A. xylabion
0.14 ± 0.01 0.67 ± 0.02
0.09
Nuclear DNA
Details of the alleles and allele frequencies at each of the three intron loci for each of A. trutta, A. truttaceus and A. georgianus were provided in Chapter 3 (Section 3.3.1). Briefly, there was a total of 19 alleles in A. trutta and eight in A. truttaceus. In addition, four alleles were found in A. xylabion (Table 4.8).
No fixed allelic differences were identified at any nDNA locus, instead the most frequent allele at each locus in each species was the same, and the alleles present in each of A. truttaceus and A. xylabion were a subset of those in A. trutta (Table 4.8). However, the results of the exact tests indicate that the frequencies of alleles in pairwise comparisons of the species samples were invariably significantly different for each locus and for all loci combined (Table 4.8).
Factorial correspondence analysis based on multilocus genotypes revealed a tendency for individuals from each species to cluster together, although there was overlap among all three species (Figure 4.4). The individuals with overlapping genotypes represented every sampling site for each species in similar proportions (data not shown). 113
Chapter 4. Molecular Taxonomy
Table 4.8
Allele frequencies at each of three intron loci for samples of three species
of Arripis and probability values for exact tests. The number of alleles sampled (n) is provided and a dash (-) indicates that the allele was not recorded in the species. The probability (p) ± standard error were determined by exact tests against a null hypothesis that allele frequencies for species in the respective pairwise comparison were the same. Pairwise comparisons, using combined data from all three loci, are provided as chisquare tests (χ2) with the relevant degrees of freedom (df). Locus/ Allele AldoB4 n 170 172 178 182 185 191 CK7 n 943 952 961 986 S72 n 816 869 892 895 900 914 919 924 930 All three loci
A. trutta
A. truttaceus
A. xylabion
182 0.005 0.011 0.775 0.011 0.011 0.187
260 0.673 0.327 -
50 1.000 -
174 0.057 0.845 0.092 0.006
74 1.000 -
40 1.000 -
168 0.006 0.006 0.006 0.006 0.143 0.565 0.238 0.024 0.006
276 0.007 0.127 0.663 0.178 0.025
42 0.762 0.238 -
A. trutta/ A. truttaceus
A. trutta/ A. xylabion
A. truttaceus/ A. xylabion
p = 0.000
p = 0.000
p = 0.000
± 0.0000
± 0.0000
± 0.0000
p = 0.000
p = 0.007
n/a
± 0.0000
± 0.0000
p = 0.011
p = 0.019
p = 0.009
± 0.0002
± 0.0003
± 0.0001
χ2 = ∞ df = 6 p = 0.000
χ2 = 35.027 df = 6 p = 0.000
χ2 = ∞ df = 4 p = 0.000
114
Chapter 4. Molecular Taxonomy
Factorial correspondence analysis (FCA) of multi-locus genotypes for
Figure 4.4
the three species of Australian Salmon. The total variation explained by each axis is given as a percentage. Arripis trutta is represented by squares (yellow); A. truttaceus by diamonds (pink); and A. xylabion by circles (orange). Note, many points overlap and are obscured.
4.4
Discussion
With multiple lines of evidence, the molecular data support the current morphologically-based taxonomy of Australian Salmon (Paulin 1993). Specifically, there was reciprocal monophyly in the mtDNA gene trees, accompanied by fixed allelic differences in mtDNA sequences in all pairwise comparisons and evidence of differences in allele frequencies in the nuclear genome.
Monophyly in the mtDNA gene trees probably represents the best single piece of evidence of the species status of the three Australian Salmon. This is because this 115
Chapter 4. Molecular Taxonomy evidence suggests that these taxa have been isolated for sufficient time for lineage sorting of these genes to have reached completion (see Avise and Ball 1990; Baum and Shaw 1995; Knowles and Carstens 2007).
However, tree-based approaches using
reciprocal monophyly for species delimitation can be limited by the fact that even highly supported clusters may not belong to a distinctive taxonomic unit, but result from chance alone (Rosenberg 2007). In order to address this, a probability test of taxonomic distinctiveness based on coalescence (a test of the null hypothesis that monophyly is a chance outcome of random branching of lineages within a species; Rosenberg 2007) was applied to each of the species of Australian Salmon. The results of these tests indicated that random branching patterns were strongly rejected as the cause of monophyly in any species (p ≤ 10-13 in all three species; data not shown). On this basis, reciprocal monophyly of the three species of Australian Salmon in the mtDNA trees almost certainly arose by evolutionary processes and provides strong support for the view that they represent true species.
Monophyly in the mtDNA sequences is also reflected in the presence of multiple fixed nucleotide differences in pairwise comparisons of the species of Australian Salmon. Such fixed differences are expected to accumulate through time, and Hey (1991) has shown that fixed differences are so unlikely to appear within an interbreeding population that the mere presence of them can be considered statistically significant.
Monophyly in the mtDNA trees, which were based on what is effectively a single locus, was not accompanied by genetic exclusivity in the nDNA data. This may reflect the point in time in the process of divergence and the fact that lineage sorting at the intron loci is incomplete. This discrepancy is likely to have arisen because the rate of genetic 116
Chapter 4. Molecular Taxonomy drift at a mtDNA locus is expected to be faster than that at a nDNA locus due to a lower effective population size in the former (Brown et al. 1979; Avise 1986; see Chapter 5). Although there was no overlap at all with mtDNA markers and it is a reasonable expectation for lineage sorting at the EPIC markers to proceed more slowly due to a larger effective population size, it is important to acknowledge the difficulty in separating ancestral polymorphism from interspecific gene flow resulting from hybridisation (introgression), especially in recently diverged species (e.g. Mallet 2005; Baack and Rieseberg 2007; Roberts et al. 2010). Nevertheless, there was evidence of allele frequency differences in the nDNA, and so evidence of genetic divergence was not limited to a single (i.e. mtDNA) genome.
As expected, the interspecific distances in the three Arripis species were substantially greater than intraspecific distances, and is taken here as part of the accumulated molecular evidence supporting the current taxonomy of Arripis. Although fixed cut-off thresholds may be useful as a ‘first step’ in delimiting species (e.g. Hebert et al. 2004; Tavares and Baker 2008; Ward 2009), it is generally recognised that a standard numerical threshold cannot be universally applied (Ferguson 2002; Moritz and Cicero 2004; Meyer and Paulay 2005; Ward et al. 2005). However, interspecific genetic distances smaller than that found in Arripis in this study have been reported between closely related species of fishes and birds (e.g. Johnson and Cicero 2004; Radchenko et al. 2008; Ward 2009).
The genetic differences between the species of Australian Salmon, although not exceptionally high, contrast with the subtle morphological differences, where few diagnostic characters exist (Paulin 1993; see Section 4.1). Paulin (1993) provided 117
Chapter 4. Molecular Taxonomy detailed morphological data for some 24 characters, of which only three differed sufficiently between the species of Australian Salmon to be useful for phenetic diagnosis. Thus, in all species of Arripis, the rate of morphological change appears to be constrained, relative to the rate of molecular evolution. In general, the process of substitution within the genome is continuous but, for many reasons, is disassociated with morphological change (Bromham et al. 2002), which may progress much slower (e.g. Inoue et al. 2005) or faster (e.g. Bromham and Penny 2003). The reasons for morphological stasis in Arripis are unknown, but may reflect functional, mechanical and physiological limitations imposed by a pelagic habit (see Barton 2007).
4.4.1
Conclusions
The molecular data in this chapter, which include information from both mitochondrial and nuclear loci, support the Arripis taxonomy of Paulin (1993).
Given that
‘populations’ can be of varying evolutionary ages and form a continuum of genetic differentiation of which ‘species’ are a part (Goldstein et al. 2000; Mallet 2007), the decision of where the population/species boundary lies remains somewhat arbitrary (Avise 2000; Hey 2006b).
Nonetheless, the three species clearly represent
independently evolving lineages (sensu de Queiroz 2007) and as such, Paulin’s (1993) “educated nomenclatural judgements” (Avise 2000, p. 308) using morphological data are now supported by multiple lines of molecular evidence.
Having established
molecular support for the current taxonomy of Australian Salmon, the next step is to investigate the relationships among these species. This investigation forms the basis of Chapter 5.
118
Chapter 5. Molecular Phylogenetics
5.0
5.1
Molecular Phylogenetics, Speciation and Biogeography
Introduction
This study has already demonstrated that: (i) each of Arripis trutta and A. truttaceus (and A. georgianus) is likely represented by a single population throughout its entire distribution in Australia (Chapter 3); and (ii) that these eastern (A. trutta) and western (A. truttaceus) populations likely represent different species (Chapter 4). However, the evolutionary relationships between these two species and a third species of Australian Salmon (A. xylabion) remain to be determined. The need to establish these relationships is inextricably linked with the development of hypotheses explaining speciation and biogeography in this group in particular and the region in general. The relationships are also important in the context of comparing the recent evolutionary histories of these closely related species, which is explored in Chapter 6.
As established in Chapter 1, A. trutta and A. truttaceus have long been considered as one example among many geminates (i.e. sister-species on either side of a barrier) in temperate Australian seas (e.g. Hutchins 1994). This is consistent with the view that the Bass Strait is a major contributor to speciation in the region, with the emergence of a land-bridge across the Bass Strait during glacial events, leading to subdivision of many faunal groups (Whitley 1932; Bennett and Pope 1953; Dartnall 1974; Womersley 1981; Hutchins 1994; O’Hara and Poore 2000). One of the most effective ways to test the (hitherto largely assumed) role of vicariance as a speciation process in this region is to assess the validity of putative geminate relationships. The presence of many east/west geminates should provide strong support for a hypothesis of vicariance (see Bowen and 119
Chapter 5. Molecular Phylogenetics Avise 1990; Lessios 2008; Rocha and Bowen 2008). However, relationships between putative geminates should not simply be assumed. This is because genetic studies have shown that, in general, presumed geminates are not always each other’s closest relatives and that their divergence may not be associated with the emergence of a barrier (e.g. Jones et al. 2003; Craig et al. 2004; Marko and Moran 2009; Miura et al. 2010). The presence of a species pair across a barrier that are not geminates does not necessarily preclude a conclusion of vicariance, especially if the species are very closely related (see Johnson and Cicero 2004; Spencer et al. 2007), but it may require a more complex explanation.
The Bassian Isthmus is an ephemeral barrier that has emerged repeatedly, raising the possibility that vicariant speciation in the region has been ongoing (see Waters et al. 2004; Rocha et al. 2005; Robertson et al. 2006; Teske et al. 2007; Corrigan and Beheregaray 2009). This can be tested by estimating the timing of speciation in a range of geminate taxa in the region, with the expectation of finding a range of divergence times among them. In this way, applying a ‘molecular clock’ (e.g. Nei and Kumar 2000; Arbogast et al. 2002; Bromham and Penny 2003) to estimate divergence times in Arripis is an important first step in assessing the degree of concordance (or not) of divergence times among species.
5.1.1
Objective
The main objective of this chapter was to generate information on the relationships among the species of Australian Salmon and use the results to develop an explanation for the speciation and current biogeography of this group. This objective will be 120
Chapter 5. Molecular Phylogenetics addressed by firstly using molecular data to establish a phylogeny for the Australian Salmon and test whether the genetic evidence supports the view that A. trutta and A. truttaceus are sister-species. Secondly, the present chapter will use a molecular dating approach to estimate the age of species divergence events within the Australian Salmon.
5.2
Materials and Methods
5.2.1
Data
The mtDNA data used in this chapter were the same as those used in Chapter 4, namely a 1137 bp concatenated sequence consisting of a 483 bp portion of the cytochrome b gene and a 654 bp portion of the cytochrome oxidase sub-unit 1 gene (COI) in Arripis. The nDNA data used in this chapter were in the form of the frequency of length variants (alleles) in the introns of four genes – the three loci used in Chapter 4 (AldoB4, CK7, S72) and an additional locus (CaM4).
Although CaM4 was monomorphic among
Australian Salmon (and thus, uninformative for determining species status in Chapter 4), it was polymorphic when the outgroup taxon (A. georgianus) was included. This locus was therefore included in this analysis because it provided greater character polarity for the calculation of a distance matrix (see Section 5.2.2.1.2).
Both the
mtDNA and nDNA data were generated using the methods described in Chapter 2. Total sample sizes for each species from samples across the entire range are provided in Table 4.2 and Table 6.2. As described in Section 4.2.3.2.1, A. georgianus was used as an outgroup for phylogenetic analyses.
121
Chapter 5. Molecular Phylogenetics 5.2.2
Data Analyses
5.2.2.1
Relationships Among Species of Australian Salmon
The summary statistics and underlying assumptions (e.g. substitution model, sequence saturation) for the concatenated mtDNA sequence data have already been determined and described in Chapter 4 (Section 4.3.1).
Neighbour-joining (NJ), maximum
parsimony (MP), maximum likelihood (ML) and Bayesian inference (BI) trees and a parsimony (haplotype) network were generated for these data as described in Chapter 4, and have already been discussed in the context of questions about species status (see Section 4.2.3.2.5). However, these trees and the network are also presented in this chapter because they also include information about the relationships among species of Australian Salmon.
5.2.2.1.1
Testing between alternative tree topologies
Three hypothetical tree topologies are possible for the three species of Australian Salmon. Specifically, the sister-species may be either: (i) A. trutta/A. truttaceus; (ii) A. trutta/A. xylabio;, or (iii) A. truttaceus/A. xylabion. The likelihoods of these competing topologies were compared by two different topology tests, as follows.
The first topology test was the Shimodaira and Hasegawa (1999) (SH) test, which used a resampling estimated log-likelihood (RELL, Kishino et al. 1990) method to compare between a posteriori selected trees. The null hypothesis for this test was that all trees were equally good explanations of the data.
Thus each of four competing tree 122
Chapter 5. Molecular Phylogenetics topologies (i.e. the three hypothetical resolved trees and a polytomy) was constrained in PAUP* version 4.0b10 (Swofford 2005), and site-wise log likelihoods calculated (see Section 4.2.3.2.5.3). The tree with the highest likelihood was identified (in this case it was the polytomy), against which each of the other trees was compared in CONSEL version 0.1i (Shimodaira and Hasegawa 2001). Confidence in each of the trees, relative to the tree with the highest likelihood, was provided as a probability value, assessed as a one-sided test (Schmidt 2009). In both this and the following test, the null hypothesis was rejected if p < 0.05.
The second topology test was the approximately unbiased (AU) test (Shimodaira 2002), based on a multi-scale RELL method. It was used here because it corrects for tree selection bias that might affect the SH method (see Shimodaira 2002).
The null
hypothesis for this test was that the expected likelihood for any one tree was greater or equal to that expected for all trees (Shimodaira 2002).
The multi-scale bootstrap
method generated replicates for many different sequence lengths, some of which were smaller and some longer than the actual sequence, and site wise log-likelihoods were calculated for each replicate as described above. Probability estimates for the AU test were calculated as the change in bootstrap support across replicates as the length of the sequence changed through the RELL method (Shimodaira 2002).
5.2.2.1.2
Intron length polymorphism
The intron allele frequency data for four loci were also used to explore the relationships among the Australian Salmon species. Specifically, a matrix of pairwise DA values (corrected chord distance; Nei et al. 1983) using allele frequency data of four intron loci 123
Chapter 5. Molecular Phylogenetics (see Section 5.2.1) in all species of Arripis was generated in DISPAN (Ota 1993). From this matrix, a 50% majority rule consensus neighbour-joining tree, based on 10,000 bootstrap replicates, was constructed with A. georgianus as an outgroup.
The corrected chord distance (DA) was used because it is the most efficient distance measure in obtaining the correct topology under the assumption of Kimura and Crow’s (1964) infinite alleles model (Nei et al. 1983; Nei and Takezaki 1994), which is considered the most appropriate model for allele frequency data (Cornuet and Luikart 1996). Assuming the mutation-drift balance is maintained, DA is expected to increase linearly with evolutionary time under the infinite alleles model (Takezaki and Nei 1996). Furthermore, DA is not biased by low-frequency alleles and low sample size (Nei et al. 1983). Neighbour-joining (Saitou and Nei 1987) was used to construct the nDNA trees because it is the most broadly applicable in this context (Nei 1991), and performs better than other methods (e.g. UPGMA) when some populations have been subject to bottlenecks (Nei and Takezaki 1994).
5.2.2.2
Exploring the Phylogenetic Signal
Given that an Arripis phylogeny was not consistently or reliably resolved (see Results), methods were applied to the mtDNA data to determine if the phylogeny was more likely to be bifurcating (tree-like) or multifurcating (polytomy) or reticulated (network-like) and to highlight possible conflicts in the bifurcating trees.
124
Chapter 5. Molecular Phylogenetics 5.2.2.2.1
Testing for a polytomy
A zero-length branch test (Slowinski 2001) was used to test whether resolving the tree was a significant improvement relative to leaving the relationship as a polytomy. This test was conducted with PAUP* using the concatenated mtDNA sequence data, where each of the three possible resolved tree topologies were treated as equally likely and thus analysed separately. One by one, branches in the tree were collapsed to zero-length and the new tree compared to the original using a likelihood ratio test (LRT; Swofford and Sullivan 2009). The LRT determined the probability that the zero-length tree was significantly worse than the starting tree (Slowinski 2001; Swofford and Sullivan 2009).
5.2.2.2.2
Likelihood mapping
A likelihood map was used to test whether the mtDNA data contained a tree-like signal (Schmidt and von Haeseler 2009) using TREE-PUZZLE version 5.2 (Schmidt et al. 2002) from all possible quartets under the TrN + G substitution model.
Maximum likelihood was used to derive posterior probabilities, p, of all possible quartets for each tree topology (a quartet is the smallest set of taxa for which more than one unrooted topology exists; Strimmer and von Haeseler 1996). The relative position of all p-values was plotted within an equilateral triangle, of which each apex represented one of the three possible tree topologies for that quartet. The resulting distribution of points within seven regions of the triangle (Figure 5.1) gives an indication of whether the phylogenetic signal is tree-like (points clustered near the apices), partly resolved network-like (points clustered along the edges) or, star-like (points clustered in the 125
Chapter 5. Molecular Phylogenetics centre). Schmidt and von Haeseler (2009) suggested that if 20% - 30% of points occur in the network-like and/or star-like areas then the phylogeny is unlikely to represent a bifurcating tree.
Figure 5.1
The
seven
regions
within
a
likelihood map. The three apices (areas 1, 2, 3) represent areas supporting strictly bifurcating trees. The sides (areas 4, 5, 6) represent areas where the decision between alternative trees is only partly resolved and the relationship might be network-like. The centre (area 7) represents points where all trees are equally supported, suggesting a star-like signal.
5.2.2.2.3
NeighborNet (Split Networks)
A NeighborNet splits graph was used to provide an assessment of possible conflict among competing bifurcating branching patterns in a phylogeny. The splits graph was generated using SPLITSTREE4 version 4.10 (Huson 1998; Huson and Bryant 2006), based on uncorrected P-distances (= Hamming) with the equal-angle algorithm. Support for edges (see below) was determined as the number of times an edge was supported in 10,000 bootstrap replicates.
126
Chapter 5. Molecular Phylogenetics NeighborNet was used in this study because it uses neighbour-joining (Saitou and Nei 1987) and least squares estimates to avoid the conservative biases of traditional split decomposition methods (Bandelt and Dress 1992; Bryant and Moulton 2004; Moulton and Huber 2009), and is one of the best network methods for assessing phylogenies that are poorly resolved by traditional methods (e.g. Kennedy et al. 2005; Morrison 2005; Wolf et al. 2007; Leaché et al. 2009; Wägele et al. 2009; Willerslev et al. 2009).
Splits are pairs of sub-trees, or bipartitions, created by the removal of a branch from a full tree and a collection of splits equates to the branches of a phylogenetic tree if every pair of splits in the collection is compatible (Moulton and Huber 2009). A NeighborNet splits graph for the Arripis data was generated as a network of compatible splits, where each branch (or edge) reflected a split with a length equal to the weight of the split (Bryant and Moulton 2004). Incompatible splits were represented as reticulations or boxes and illustrated that there were several possible paths between taxa. Thus, each split in the splits graph corresponded to a collection of parallel edges, all with the same length. Empirical support was provided by bootstrapping (Huson and Bryant 2006).
5.2.2.3
Molecular Dating
5.2.2.3.1
Testing for a clock-like rate of substitution
In order to determine the appropriate method for estimating divergence times, it was necessary to test for a uniform rate of nucleotide substitution in the concatenated mtDNA sequences of all Arripis lineages. This was done using a likelihood ratio test (LRT) for the assumption of a molecular clock (Felsenstein 1981; Huelsenbeck and 127
Chapter 5. Molecular Phylogenetics Crandall 1997).
The test was conducted in TREE-PUZZLE using the TrN + G
substitution model and A. georgianus as a rooted outgroup. Using likelihood, the LRT compared the goodness-of-fit between two models: (i) an enforced molecular clock (allowing only a single rate of nucleotide substitution among all lineages); and (ii) a phylogeny that was not constrained to a single rate of substitution, which theoretically represented a more complex model (i.e. branches could evolve at different rates). For the Arripis sequence data, the assumption of a molecular clock was not rejected because the likelihood of the more complex (no clock) tree was not significantly higher (∆ = 12.64, df = 20, p = 0.8924).
5.2.2.3.2
Divergence rate
In this chapter, a rate of 2% sequence divergence per million years between lineages (Brown et al. 1979) was used in order to estimate divergence times between the species of Arripis. This rate was used because it is widely applied to mitochondrial genes (including cytochrome b and COI) as a ‘standard’ rate when no other accurate estimates (from fossils or related extant taxa) are possible (e.g. Burridge 2002; Theisen et al. 2008; Drew et al. 2009), as was the case for Arripis (see Discussion).
5.2.2.3.3
Time to most recent common ancestor (Tmrca)
Molecular dating was performed on the concatenated mtDNA sequence data using Bayesian Markov Chain Monte Carlo (MCMC) resampling with the program BEAST version 1.4.8 (Drummond and Rambaut 2007). Based on the results of the tests for a uniform rate of substitution across all lineages (see Section 5.2.2.3.1), a strict clock was 128
Chapter 5. Molecular Phylogenetics specified. A Yule speciation process was specified as a tree prior along with model parameters derived from JMODELTEST (see Table 4.3). The MCMC analyses were run for 10 million generations sampling every 100 generations with the first 10% discarded as burn-in.
Given limited support for one topology over another (see
Results), each of the three possible tree topologies (representing three different sisterspecies relationships among the Australian Salmon) was analysed separately. In each analysis, a topology was constrained so the MCMC search was limited to trees satisfying a particular sister-species relationship between the three species of Australian Salmon.
Results were inspected using TRACER version 1.4.1 (Rambaut and
Drummond 2007) to confirm that stationarity and adequate effective sample sizes had been obtained for all parameters indicating that the analysis had run for a sufficient number of generations. Confidence was provided by 95% highest posterior density (HPD) intervals, which is a credible set of trees containing 95% of the sampled values (Drummond and Rambaut 2007).
5.3
Results
As discussed in Chapter 4, the nucleotide composition of a 1137 bp concatenated mtDNA sequence was determined for 25, 36 and 11 individuals of A. trutta, A. truttaceus and A. xylabion, respectively, and revealed the presence of 14, five and two haplotypes, respectively, and 32 polymorphic sites among species (Table 4.4). Arripis trutta sequences contained five fixed sites, A. truttaceus four and no completely fixed differences were found in A. xylabion (Table 4.5), although there were fixed differences in this species in pairwise comparisons (see Section 4.3.1).
129
Chapter 5. Molecular Phylogenetics 5.3.1
Relationships Among Species of Australian Salmon
The phylogenetic trees and the parsimony network did not provide consistent support for a sister-species relationship between A. trutta and A. truttaceus (Figures 5.2, 5.3). For the mtDNA data, all phylogenetic (tree-building) methods resulted in either unresolved topologies or resolved topologies with very low support for interspecific relationships (Figure 5.2; Table 5.1).
In three of the four methods (maximum
parsimony, maximum likelihood, Bayesian inference), the relationship was represented as a trifurcating polytomy with each species arising from a single node (Figure 5.2B, C). Only neighbour-joining reconstructed a resolved consensus tree (with A. trutta and A. truttaceus as sister-species), although bootstrap support was very low (Figure 5.2A). Note that similar results were also found using each gene region independently in that, even if a tree was resolved, bootstrap support was low (Table 5.1, details not shown).
The neighbour-joining phylogenetic tree constructed using the nDNA allele frequency data indicated a sister-species relationship between A. truttaceus and A. xylabion, however, bootstrap support was very low (Figure 5.4).
130
Chapter 5. Molecular Phylogenetics
Figure 5.2
(over page) Majority consensus phylogenies of Australian Salmon
inferred from concatenated mtDNA sequence haplotypes based on four methods. A. Neighbour joining (NJ) method using distances based on the TrN + G model. B. Maximum parsimony (MP; bootstrap values above line), utilising a heuristic search with TBR branch swapping. Bayesian inference (BI), using the GTR + G model generated the same tree (posterior probabilities below line). C. Maximum likelihood (ML), using the TrN + G model with a heuristic search and TBR branch swapping. Clades are colour coded by species. Haplotypes are numbered as in Table 4.5. Node support values were derived from 1000 bootstrap replicates (NJ, MP, ML) or 15,000 sampled post burn-in trees (BI). In each tree, the congener Arripis georgianus was the outgroup. 131
Chapter 5. Molecular Phylogenetics
Figure 5.3
Parsimony network for the mtDNA haplotypes from the concatenated
sequence data (cytochrome b and COI) of the three species of Australian Salmon obtained by the criterion of parsimony with the program TCS version 1.21 (Clement et al. 2000). Each line in the network shows a single mutational change. A haplotype is shown by a coloured circle, the surface area of which is proportional to the frequency of the haplotype. Intermediate, unsampled haplotypes are indicated by a small black dot. Similar networks are shown for each of the genes separately (inset). In each case, Arripis trutta is represented by yellow; A. truttaceus by pink; and A. xylabion by orange.
133
Chapter 5. Molecular Phylogenetics Summary of sister-species relationships between the three species of
Table 5.1
Australian Salmon generated by four phylogenetic methods with the concatenated mtDNA sequence data. Methods are neighbour-joining (NJ), maximum parsimony (MP), maximum likelihood (ML), and Bayesian inference (BI). Where a relationship was resolved, the sister-species are listed with either bootstrap or Bayesian support in parentheses. Unresolved relationships are indicated as a polytomy. See Figure 5.2. concatenated
cytochrome b
COI
NJ
A. trutta A. truttaceus (52)
A. truttaceus A. xylabion (52)
A. trutta A. xylabion (61)
MP
polytomy
A. truttaceus A. xylabion (59)
polytomy
ML
polytomy
A. truttaceus A. xylabion (55)
polytomy
BI
polytomy
A. truttaceus A. xylabion (0.69)
polytomy
Figure 5.4
Majority consensus phylogeny of Australian Salmon inferred from a DA
distance matrix of allele frequency data from length polymorphism in four intron loci, constructed by neighbour-joining. Node support value is derived from 1000 bootstrap replicates.
Connections to the outgroup have been shortened for scaling purposes
(dotted line). 134
Chapter 5. Molecular Phylogenetics 5.3.1.1
Support for Alternative Tree Topologies
The results of both the SH and AU topology tests, which were applied to the mtDNA data in view of the absence of compelling support for any particular tree topology, indicated that the likelihood of all possible resolved tree topologies were not significantly different from each other or from a polytomy (Table 5.2).
Table 5.2
Probability values derived from two topology tests for support of each of
the possible resolved tree topologies among the three species of Australian Salmon using the concatenated mtDNA sequence data.
Probability for the SH test was
determined by comparing the constrained tree against the highest likelihood tree (a polytomy; see Section 5.2.2.1.1). Probability for the AU test equates to multi-scale bootstrap support for the null hypothesis (Ho). Null hypotheses are rejected when p < 0.05. Note that each topology was an independent test.
* †
Shimodaira & Hasegawa (SH)*
0.828
0.172
0.172
Approximately Unbiased (AU)†
0.826
0.174
0.176
Ho (SH) = all trees are equally likely Ho (AU) = likelihood for constrained tree is greater or equal to that expected for all bootstrap trees
135
Chapter 5. Molecular Phylogenetics 5.3.2
Exploring the Phylogenetic Signal
The three approaches that were used to further explore the phylogenetic signal in the mtDNA data (i.e. zero-length branch test, likelihood mapping, NeighborNet), failed to find support for a bifurcating tree.
The zero-length branch tests were unable to reject the null hypothesis of a polytomy (i.e. a topology with zero-length branches) for any of the three possible bifurcating tree topologies. Specifically, none of the trees with bifurcating topologies generated a tree with an improved likelihood over a polytomy (all p-values = 1.00).
Likelihood mapping also indicated that it is not possible to find a resolved bifurcating tree with the mtDNA data. Specifically, based on the guidelines of Schmidt and von Haeseler (2009) (see Section 5.2.2.2.2), the analysis revealed considerable evidence for a non-bifurcating signal (areas 4 + 5 + 6 + 7 ≈ 49% of points; Figure 5.5) with only weak support for a tree-like signal (areas 1 + 2 + 3 ≈ 51% of points; Figure 5.5).
In the NeighborNet splits graph, there was considerable bootstrap support for each of the species ‘clades’ (81% - 99%) and hence a high least squares fit (99.9) (Figure 5.6). However, the graph contained conflicting signals at the point of divergence of the three species of Australian Salmon, as indicated by the presence of parallelograms, such that the maximum level of bootstrap support for any of the alternative pathways was only 34.4% (Figure 5.6).
136
Chapter 5. Molecular Phylogenetics
Figure 5.5
Likelihood map for concatenated mtDNA sequence data for the three
species of Australian Salmon. A. Likelihood map indicating the percentage of points contained within each of seven regions. B. Point map showing relative position of posterior probabilities used to generate the percentage values in A. The apices represent support for bifurcating trees; the edges, support for network-like trees and; the centre, support for star-like trees (see Section 5.2.2.2.2).
Figure 5.6
(over page) NeighborNet splits graph based on uncorrected P-distances
with the rooted equal-angle algorithm on the concatenated mtDNA sequence data for the three species of Australian Salmon. Arripis georgianus was the outgroup root (not shown). Haplotypes are indicated by Roman numerals as given in Table 4.5. Parallel edges (branches) represent alternative pathways between taxa. Bootstrap support is indicated for the edges leading to the ‘clades’ representing each species (all > 80%) and for the edges indicating alternative basal pathways between the three species (all < 35%; enlarged inset). P-distances are indicated by the scale bar. 137
Chapter 5. Molecular Phylogenetics
138
Chapter 5. Molecular Phylogenetics 5.3.3
Molecular Dating
Phylogenetic trees based on the assumption of a molecular clock indicated that branch lengths between the Australian Salmon species were very short, irrespective of which of the three possible topologies was enforced, and all topologies had similar posterior likelihood probabilities (Table 5.3). The mean estimates of Tmrca for the divergence of each of the three Australian Salmon were very similar, regardless of the tree topology, and suggested a Pleistocene divergence some 647,000 – 704,000 years ago (see Node ‘a’ in Table 5.3). Confidence intervals of these estimates ranged from 249,000 years to 1,060,000 years and overlapped across all tree topologies and also between nodes (Table 5.3), although they all inevitably place the divergence in the mid to late Pleistocene.
139
Table 5.3
Estimates of the time to most recent common ancestor (Tmrca) for each of the three possible tree topologies for Australian
Salmon based on Bayesian resampling of the concatenated mtDNA sequence data using a divergence rate of 2% per million years. Tmrca estimates and confidence (95% highest posterior density interval; in parentheses) are given in millions of years before present. Node ‘a’ and ‘b’ refer to the respective node annotated on each pictogram. The estimated tree height is equivalent to an estimate time since divergence from the outgroup, Arripis georgianus. The Bayesian posterior likelihood for each possible tree topology is provided.
Posterior likelihood
-2259.309
-2303.929
-2259.483
Tree height
7.830 (7.305 – 13.459)
13.516 (9.275 – 23.230)
8.002 (7.473 – 14.226)
Tmrca (Node a)
0.675 (0.451 – 0.945)
0.704 (0.441 – 0.968)
0.647 (0.493 – 1.060)
Tmrca (Node b)
0.669 (0.430 – 0.888)
0.668 (0.383 – 0.854)
0.507 (0.249 – 0.750)
Chapter 5. Molecular Phylogenetics 5.4
Discussion
The molecular data in this chapter provided no definitive support for the geminate status of the Arripis species on the east and west coasts of Australia, A. trutta and A. truttaceus, respectively. In fact, the evidence is consistent with a trifurcating polytomy where the three species of Australian Salmon (the two mentioned above and A. xylabion) arose at, or nearly at, the same time.
5.4.1
Hard Polytomy in Australian Salmon
The molecular phylogeny of the three species of Australian Salmon is probably a ‘hard’ polytomy (e.g. Coddington and Scharff 1996), and likely reflects a radiation, in which the three species arose from a common ancestor at, or almost at, the same time. However, distinguishing a hard polytomy from a soft polytomy (reflecting systematic errors or low resolution data) is difficult. Although the default has often been to assume a soft polytomy, this rejects a polytomy as a valid and likely phylogenetic scenario (Coddington and Scharff 1996; Page and Holmes 1998) and examples of hard (or effectively hard) polytomies with short internal branch lengths are being increasingly recognised (e.g. Poe and Chubb 2004; Barth et al. 2008).
The polytomy might represent a true trifurcating radiation, in which three Australian Salmon species arose at precisely the same time. More likely, however, is that the lineages diverged over a short time period, relative to the rate of character evolution, and so internal branches did not have time to accumulate mutations (e.g. Walsh et al. 1999; Braun and Kimball 2001; Whitfield and Lockhart 2007). If this is the case, there 141
Chapter 5. Molecular Phylogenetics probably is insufficient evolutionary information in the genome to be able to reliably reconstruct the precise order of divergences, in which case the polytomy is ‘effectively hard’ (see Hoelzer and Melnick 1994; Page and Holmes 1998). This is supported by a maximum-likelihood power analysis (Walsh et al. 1999), which indicated that, based on the detected levels of polymorphism, even sequencing the entire Arripis mtDNA genome would permit detection of two divergences only if they occurred at least 30,000 years apart (data not shown). The same power analysis method also indicated that > 117,000 bp (more than six times the length of the entire mtDNA) would be required to discriminate between divergences that occurred within 10,000 years of each other.
On the assumption that the mtDNA gene tree is a hard polytomy, and is consistent with the species tree (e.g. Maddison 1997), the results of this chapter infer that A. trutta and A. truttaceus are not, by definition, true geminates (sensu Jordan 1908). While this has potential implications for our understanding of speciation in southern Australia (see below), the eastern and western species of Australian Salmon may nevertheless be considered as ‘sibling’ species (Jordan 1908; Johnson and Cicero 2004). As such, they still provide an excellent opportunity for comparing the recent evolutionary histories of species on the east and west coasts of Australia (see Chapter 6).
Despite compelling evidence for a hard polytomy, it remains difficult to completely rule out a soft polytomy. In theory, a soft polytomy could have been caused by low levels of polymorphism in the mtDNA sequences, a systematic error in the data (e.g. from the concatenation), and/or inadequate sampling (see Coddington and Scharff 1996; Slowinski 2001).
The amount of polymorphism in the cytochrome b and COI
sequences of Arripis was low, despite these being among the most phylogenetically 142
Chapter 5. Molecular Phylogenetics useful and widely used (see Avise 2004; Hebert et al. 2004; see also Chapters 1 and 2). To a certain extent, this problem was resolved by concatenating the two types of sequence. Although concatenation of sequences can lead to errors in phylogenetic analyses (e.g. Huelsenbeck et al. 1996), this is unlikely to be a problem in the present study because the combining of the cytochrome b and COI sequences was validated by the partition-homogeneity test (see Section 4.2.3.2.2). Inadequate sampling also does not appear to have been a major problem in the present study. This is because the analyses included at least 25 individuals of each of A. trutta and A. truttaceus from across the entire geographic range of the species. The sampling range for A. xylabion was more limited as it was not possible to obtain samples of this species from any site other than Lord Howe Island. Furthermore, for all three species of Australian Salmon, sample size increases during the production of the mtDNA data only resulted in detection of more rare haplotypes, one or two mutational steps away from the common haplotype (see parsimony networks), which offer little phylogenetic utility (see Nei and Kumar 2000). Increasing the number of genes sampled (both mtDNA and nDNA) is probably required to provide greater confidence in a hard polytomy (e.g. Kliman et al. 2000; Waddell et al. 2000; Cooper et al. 2001; Kraus et al. 2006; Willerslev et al. 2009). A genome-wide, species tree approach (Maddison 1997; Brito and Edwards 2008; Avise 2009; Edwards 2009) would provide the ultimate genetic test of the relationships in Arripis, however if the divergences were close in evolutionary time as predicted from the current data set, this probably would not fundamentally change the outcome – i.e. an effectively hard polytomy.
143
Chapter 5. Molecular Phylogenetics 5.4.2
Pleistocene Divergence
Regardless of the relationships among the three species of Australian Salmon, the mtDNA results of the present chapter indicate that these species probably diverged from the most recent common ancestor in the mid to late Pleistocene, some 700,000 years ago. These estimates of divergence time are similar to allozyme-based estimates for A. trutta and A. truttaceus (MacDonald 1980) in that the divergence is placed in the mid to late Pleistocene and the confidence intervals overlap, although the allozyme estimates are slightly earlier (0.8 – 1.14 Mya). Although the divergence estimates of the present study are based on the standard mtDNA substitution rate of 2%/Myr (which might not be ideal), the consistency of divergence estimates between the two studies suggests that this rate provides a reasonable approximation for Arripis (see also Grant and Bowen 1998; Avise 2000; Burridge 2002; Theisen et al. 2008; Weir and Schluter 2008; Drew et al. 2009). Regardless, in the case of the Arripidae, it is impossible to obtain precise data on the rate of molecular evolution because there are no fossil records for the family (M. Siversson, Western Australian Museum, pers. comm.) and even the phylogenetic position of the family is uncertain (K. Carpenter, Old Dominion University, pers. comm.) (see Bromham and Penny 2003). Finally, while the data suggest a mid to late Pleistocene divergence, it is important to recognise that the estimated divergence times for Arripis are crude (e.g. have large confidence intervals) such that it is impossible to directly link them to specific glacial events within the Pleistocene.
In the present study, even the most conservative estimate of divergence time for the three Australian Salmon (1.05 Mya; see Table 5.3: HDP, Node ‘a’), places the split in the mid Pleistocene. Notwithstanding the difficulties of direct comparisons across 144
Chapter 5. Molecular Phylogenetics studies, this is later than available estimates for other geminate marine species in temperate Australia. For example, using a similar type of data, Burridge (2000b) estimated that two pairs of teleost sister-species diverged well into the Miocene (7.8 – 11.2 Mya) and another in the Pliocene (2.7 – 3.8 Mya), while divergences between east/west sister-species of sharks (Corrigan and Beheregaray 2009) and sea-stars (Waters et al. 2004) were calculated as near the onset of glaciation around the Pliocene/early Pleistocene (1.0 – 3.9 Mya, and 2.1 – 2.4 Mya, respectively). At the other extreme, the presence of divergent (conspecific) populations across the Bass Strait in a range of taxa provides circumstantial evidence of much more recent divergence (e.g. Dawson 2005; Waters et al. 2005; York et al. 2008; Fraser et al. 2009). It is likely that at least some of these east/west populations will eventually evolve into separate species (or be recognised as such; see Spencer et al. 2007), and thus become true geminates. The apparently wide range of divergence times for pairs of marine taxa on either side of the Bass Strait is consistent with the view that the (ephemeral) Bassian Isthmus has acted as an isolating barrier on multiple occasions, both prior to and throughout the Pleistocene and into the present. Similar conclusions have been reached for other temporary marine barriers (Bowen and Avise 1990; Bernardi et al. 2003).
5.4.3
Speciation in Australian Salmon
While a simple hypothesis invoking the Bass Strait as an isolating mechanism among sister-species in temperate Australian marine fauna may be appealing (see Section 5.1), the results of this chapter suggest that the Bassian Isthmus might only be one contributing factor to species radiations in the region. As sibling species (see above), the relationship between A. trutta and A. truttaceus is still consistent with speciation by 145
Chapter 5. Molecular Phylogenetics vicariance associated with the Bassian Isthmus and oceanographic conditions (see Chapter 1). Specifically, the emergence of a continuous land bridge across the Bass Strait, in conjunction with colder sea temperatures (Wells and Okada 1996; see also Chapter 6), could be expected to have vicariantly isolated populations and caused them to move northwards to escape the cold water. Northward movement of individuals on the west of the Bassian Isthmus would inevitably have taken these individuals on a considerable journey around the south-west corner and onto the west coast of Australia. This most likely would have involved adult fish since the west to east direction of prevailing winds and currents at that time suggests that passive dispersal by larvae from east to west was/is very unlikely (CLIMAP 1976; Prell et al. 1980; Petrusevics and Bye 1995). A similar scenario was proposed for two other east/west sister-species of fishes (in the genera Chironemus and Nemadactylus; Burridge 2000a;2000b).
In contrast to the species on the east and west coasts of Australia, the presence of another sibling-species of Arripis at Lord Howe Island and nearby islands is probably best explained as allopatric speciation by peripheral isolation. During glacial periods, in addition to the presence of a land bridge across the Bass Strait, current flows would have likely facilitated the dispersal of larvae of Arripis (and other species, e.g. Waters and Roy 2003) from the east coast of Australia to the islands north of New Zealand. Specifically, eddies diverging from the mainland were thought to be positioned further north, thus a branch of water, carrying larval Arripis, could have reached Lord Howe Island (31.5o S; Bostock et al. 2006; Woodroffe et al. 2006). Geomorphological cores around Lord Howe Island show that the current tropical community is Holocene in origin (< 7000 years old) and the island was dominated by temperate marine communities throughout much of the Pleistocene (Woodroffe et al. 2006), indicating the 146
Chapter 5. Molecular Phylogenetics availability of suitable conditions when the species of Arripis are predicted to have diverged. In a related way, this hypothesis also provides an explanation for why A. trutta is not found on Lord Howe Island, but is found in New Zealand. In present conditions, the East Australian Current diverges from the mainland at around 32o S (Bostock et al. 2006) and flows to the south of Lord Howe Island, thus larval A. trutta are transported to New Zealand, rather than to Lord Howe Island. This is reflected in contemporary dispersal to Lord Howe Island, which is found in tropical taxa with sources north of 32o S (e.g. mangroves, Duke et al. 1998; corals, Ayre and Hughes 2004; sea turtles, Boyle et al. 2009), rather than temperate fauna such as A. trutta.
5.4.4
Conclusions
Phylogenetic relationships among the three species of Australian Salmon were not fully resolved despite the use of a range of analytical approaches. It is suggested that this is because the species arose within a very short timeframe and consequently probably represent an effectively hard polytomy. As such, a ‘resolved’ bifurcating phylogeny may not exist and by definition, true ‘sister’ species may not exist. Rather, the three species might best be considered ‘sibling’ species that appear to have diverged in the mid to late Pleistocene, which is later than the divergence times estimated for other geminate species in the same region. The available evidence is consistent with the view that glaciation may have caused a species radiation in Arripis by two different modes – vicariance, leading to the east and west coast species (A. trutta and A. truttaceus, respectively) and dispersal to the east (A. xylabion). This scenario also provides for a relatively fast radiation, since a single glacial cycle would simultaneously result in the conditions that could lead to both mechanisms. 147
Chapter 6. Recent Evolutionary History
6.0
6.1
Genetic Diversity and Recent Evolutionary History
Introduction
As explained in Chapter 1, the presence of putative geminate species in southern Australia provides an opportunity to explore the evolutionary processes affecting the fauna in the region. Although the east/west species pair of A. trutta and A. truttaceus were not confirmed as true sister-species, it is nevertheless likely that they are part of a trio of sibling-species that arose nearly simultaneously (see Chapter 5). Consequently, this pair can effectively be regarded as geminates in such explorations. In particular, these two species can be used to test whether glacial cycles and associated changes in environmental conditions have impacted more severely on species distributed on the western side of the Bass Strait compared to the eastern side (e.g. Fraser et al. 2009).
As discussed in Chapter 5, the species of Australian Salmon (A. trutta, A. truttaceus, A. xylabion) likely arose during the mid to late Pleistocene (~ 700,000 years ago). Since then, the climate and oceanography of southern Australia (and other locations) has experienced marked fluctuations through numerous glacial cycles (Jansen et al. 2007). These glacial cycles have had a significant impact on coastal marine environments in the region and, associated with this, can be expected to have had a significant impact on the demographies and distributions of the resident flora and fauna (e.g. Sotka et al. 2005; Janko et al. 2007; Fraser et al. 2009).
Interestingly, a range of historical
environmental data indicate that the environmental changes accompanying glacial cycles have been more severe on the western side of the Bass Strait compared to the eastern side, as is explained below. Consequently, it can also be expected that the 148
Chapter 6. Recent Evolutionary History demography and distribution of species such as A. truttaceus on the western side will potentially be more severely impacted by glacial cycles than species such as A. trutta on the eastern side.
Although there have been numerous glacial cycles since the mid to late Pleistocene, the Last Glacial Maximum (LGM; ~20,000 years ago) is the best known. As a result, the details of the LGM are often used as a proxy for interpreting earlier events (Paulo et al. 2001; Jansen et al. 2007), despite the fact that some aspects (e.g. magnitude and duration) of the LGM might have been different. In the context of genetic studies, the LGM is particularly important because its effects, if strong enough, may erase signals from earlier events (Hewitt 2000).
6.1.1
Pleistocene Glacial/Interglacial Cycles in Southern Australia
6.1.1.1
East of the Bass Strait
Both sea temperatures and current flows on the east coast of Australia differed only slightly between glacial and interglacial stages (CLIMAP 1976; Lawrence and Herbert 2005; Hostetler et al. 2006). For example, during the LGM, the water in the Tasman Sea was probably only substantially colder in the southern-most parts (CLIMAP 1976; Hostetler et al. 2006). Consequently, for example, glacial and interglacial foraminiferal assemblages on the east coast were similar, differing mostly in subtle latitudinal shifts (Bostock et al. 2006).
The similarity of water temperatures between glacial and
interglacial stages was largely due to two factors, as follows.
149
Chapter 6. Recent Evolutionary History Firstly, during the LGM, east of the Bassian/Tasmanian peninsula, the position of the Subtropical Convergence Zone (which is the interface between cool southerly waters and warmer, tropically derived waters) traced the southern boundary of the Tasman Sea between Tasmania and New Zealand.
This is only slightly north of the present
(interglacial) position.
Secondly, the southward flow of warm water via the East Australian Current, which dominates the patterns of present-day water movement down the east coast, continued through the LGM, albeit reduced in strength and slightly cooler (CLIMAP 1976; Bostock et al. 2006; Hostetler et al. 2006).
However, the point where the East
Australian Current diverts from the mainland towards New Zealand (the Tasman Front) was pushed a few degrees northwards (Bostock et al. 2006; see also Chapter 5). Nevertheless, throughout the LGM, an extension of the East Australian Current continued south of the Tasman Front to Tasmania (Bostock et al. 2006; Figure 6.1).
Figure 6.1
(over page) Circulation features of southern Australia (modified from
Wells and Okada 1996). A. Present (interglacial) conditions: Subtropical Convergence Zone is positioned south of the continent. The south flowing Leeuwin Current carries warm water down the west coast and along the south coast. The cool, offshore Western Australian Current flows northwards. B. Reconstruction of Late Quaternary glacial conditions: Polar ice-sheet expanded, lowering sea levels and exposing the Bassian Isthmus. The Subtropical Convergence Zone shifted northwards. The Leeuwin Current ceased and the Western Australian Current dominated on the west coast.
150
Chapter 6. Recent Evolutionary History
151
Chapter 6. Recent Evolutionary History 6.1.1.2
West of the Bass Strait
In contrast to the eastern side of the Bass Strait, both sea temperatures and current flows on the western side differed markedly between glacial and interglacial stages (CLIMAP 1976; Lawrence and Herbert 2005; Hostetler et al. 2006).
For example, sea
temperatures were likely to have been significantly lower than the present - around 4o C less on the west coast and up to 10o C lower on the south coast. This was largely due to three factors, as follows.
Firstly, during the LGM, the Subtropical Convergence Zone shifted northwards by several degrees (due to an expansion of the Antarctic ice-shelf) and was juxtaposed with the southern Australian coastline from the south-west corner to the exposed Bassian Isthmus (CLIMAP 1976; Howard and Prell 1992; Wells and Okada 1996; Figure 6.1).
Secondly, circulation patterns in the southern Indian Ocean were deflected off the Subtropical Convergence Zone causing an exaggerated flow of cool water to the southwest of Australia, which was diverted northwards by the continental land mass (CLIMAP 1976; Prell et al. 1980). This flow of cool water corresponds to the Western Australian Current, which, while weak during interglacial stages (including the present), dominated the circulation along the west coast of Australia during glacial stages (Belyaeva and Burmistrova 1992; Figure 6.1).
Thirdly, and partly as a consequence of the strengthened northward flow of the Western Australian Current, the Leeuwin Current, which dominates the patterns of present-day winter water movement down the west coast and along the south coast (Godfrey and 152
Chapter 6. Recent Evolutionary History Ridgway 1985), completely ceased during glacial stages (Wells and Wells 1994; Gingele et al. 2001; Figure 6.1).
6.1.2
Using Genetic Data to Infer the Evolutionary Histories of Arripis Species
Even when fossil and sub-fossil records exist, it is typically difficult to obtain good information about the evolutionary history of a species and, in most cases, genetic data provide the only option for investigating demographic history, as is the case for Arripis. Genetic approaches in this regard are part of the rapidly expanding field of phylogeography (e.g. Avise 2000;2009) and hinge on the fact that patterns in genetic data, especially departures from mutation-drift equilibrium, can provide clues about the recent demographic history of a species, such as whether the size of a population has been stable, growing or contracting during certain time periods (Watterson 1984; Grant and Bowen 1998; Avise 2000; Knowles 2009).
Nevertheless, it is important to
recognise that the interpretation of genetic data in this context is not always straight forward, and is often dependent on restrictive assumptions (e.g. see Avise 2000; Hare 2001).
6.1.3
Aim
The main aim of this chapter was to assess and compare the levels and patterns of genetic diversity in the effectively geminate species A. trutta and A. truttaceus. The resultant information was used to infer and compare aspects of recent evolutionary histories of these species.
In particular, this chapter tests the prediction that A.
truttaceus, which resides to the west of the Bass Strait, may have been more severely 153
Chapter 6. Recent Evolutionary History impacted by glacial cycles than A. trutta, which resides to the east. Finally, data on A. georgianus, which is sympatric with A. truttaceus, and on A. xylabion on Lord Howe Island, are also presented to provide context for the results for Eastern and Western Australian Salmon.
6.2
Materials and Methods
6.2.1
Sampling Regime
The results of this chapter are based upon the analysis of samples of A. trutta, A. truttaceus and A. georgianus from across their Australian distributions, and of A. xylabion from Lord Howe Island. Since one of the main aims of this chapter was to explore the effects of glaciation on populations to the east and west of the Bass Strait in Australian waters, and since the relationship between the Australian and New Zealand assemblages of A. trutta is not clear, the two individuals of this species from New Zealand were not included. [NB: although, the exclusion of these individuals had a negligible effect on all results; data not shown]. The sample sizes for each of the sampled mtDNA and intron loci for each sampling site for each species are provided in Chapter 4 (Table 4.2). Given no evidence of population structure in any species in Australian waters (Chapter 3), all samples were pooled into a single sample for each species as described in Chapter 4.
154
Chapter 6. Recent Evolutionary History 6.2.2
Genetic Assays
The mitochondrial DNA data used in this chapter are in the form of the nucleotide sequence of a 483 bp portion of the cytochrome b gene in Arripis, generated as described in Chapter 2. The cytochrome b region was selected for mtDNA analyses because it contained a higher level of polymorphism than the COI region. Concatenated sequences (i.e. cytochrome b plus COI) were not used in this chapter because, while concatenation may improve phylogenetic resolution between species (see Section 4.2.3.2.2), they may impose a greater risk of error when used for demographic inference (Brito and Edwards 2008; Edwards 2009; Knowles 2009).
The nuclear DNA data used in this chapter are in the form of the frequency of length variants (alleles) in the introns of four genes (AldoB4, S72, CaM4 and CK7). These loci were selected for study because, of the 12 intron loci that could be reliably amplified in Arripis (see Chapters 2 and 3), only these four loci had scorable length polymorphism in one or more species of Arripis.
However, one measure of genetic diversity (the
proportion of polymorphic loci) was based on data from all 12 of the scorable loci (see Chapter 2). Alleles were binned into size classes as explained in Section 2.5.4. The nDNA data were generated using the methods described in Chapter 2.
6.2.3
Genealogical Relationships
The evolutionary relationships among the cytochrome b haplotypes of A. trutta, A. truttaceus and A. xylabion are discussed in Chapters 4 and 5 in the context of species relationships. The associated parsimony networks (excluding the two New Zealand 155
Chapter 6. Recent Evolutionary History specimens) are repeated in this chapter, along with the equivalent parsimony network from a sample of 30 individuals of A. georgianus (presented for the first time), because they can also provide circumstantial evidence in support of demographic inferences (e.g. Avise 2000; Mateus et al. 2011).
6.2.4
Data Analyses
6.2.4.1
Genetic Diversity
For the mtDNA data, genetic diversity was measured in three ways: (i) the standardised number of haplotypes; (ii) haplotype diversity; and (iii) nucleotide diversity.
The standardised number of haplotypes (SNH) was calculated using POPTOOLS (www.cse.csiro.au/CDG/poptools).
Since the number of haplotypes observed in a
sample strongly depends on the size of the sample, estimates of the number of haplotypes were standardised according to sample size in order to compare between species. The standardising procedure was based around the number of individuals present in the smallest sample (i.e. 23 for A. trutta). Thus, for each of A. truttaceus and A. georgianus, 23 individuals were shuffled (i.e. sampled without replacement) over 1000 times. This method was used because simulations have shown that sampling without replacement is one of the most robust and precise methods for standardising haplotype numbers (Leberg 2002).
Estimates of haplotype diversity (and standard errors) were calculated using ARLEQUIN version 3.1 (Excoffier et al. 2005).
Haplotype diversity (h) is the 156
Chapter 6. Recent Evolutionary History probability that two haplotypes, randomly chosen from a sample, are different (Nei 1987).
Estimates of nucleotide diversity were also calculated using ARLEQUIN.
Nucleotide diversity (π) is the probability that two homologous nucleotides, randomly selected from a sample, are different (Nei 1987). It was estimated according to the number of substitutions between two sequences, using the substitution model TrN (see Section 4.2.3.2.4 for details of the choice of model) with a gamma correction (G) of 0.02 in A. trutta and 97.03 in both A. truttaceus and A. georgianus.
For the nDNA data, genetic diversity was measured in three ways: (i) proportion of polymorphic loci; (ii) standardised number of alleles; and (iii) average expected heterozygosity per polymorphic locus. The proportion of polymorphic loci (PO) was calculated as the proportion of the 12 intron loci where the common allele occurred with a frequency less than 0.95.
The standardised number of alleles (SNA) was calculated for each polymorphic locus for each species and then the values were summed over all loci to provide a standardised estimate of the total number of alleles for each species. As described above, for each locus, samples for A. truttaceus and A. georgianus were standardised to the number of A. trutta sampled (the species with the smallest n), by shuffling over 1000 times in POPTOOLS.
Expected heterozygosity (HE) in all species was averaged for four loci (AldoB4, CaM4, CK7, S72). Average expected heterozygosity and its variance was calculated using the method of Nei (1987), as implemented in ARLEQUIN.
157
Chapter 6. Recent Evolutionary History 6.2.4.2
Comparisons of Genetic Diversity Between Species
For each of the five sampled loci, each species was ranked from highest to lowest according to the amount of diversity present and trends in the data were inspected in order to assess whether the rankings of the species were consistent across loci. This was done separately for values of: (i) the standardised number of haplotypes/alleles; and (ii) haplotype diversity/expected heterozygosity. Simple ranks were used (e.g. Kendall 1970) and an average of the ranks was used when values were tied between two or more species (e.g. Legendre 2005). It was not possible to assess the statistical significance of these results because, for the relevant test, Kendall's coefficient of concordance for ranks test (W; Kendall 1970), there were too few polymorphic loci in Arripis (see Legendre 2005; Chang 2009).
6.2.4.3
Neutrality of Markers
The neutrality of the markers in each species of Arripis was assessed because whether or not the markers are under the influence of selection has a major influence on the interpretation of the results. Kendall's coefficient of concordance for ranks test (W) was used to test for evidence of neutrality via concordance in the patterns of variation at different loci, based on the theoretical assumption that, if loci are essentially selectively neutral, then concordant patterns of genetic diversity should be detected across all loci (Nei 1987; Avise 2004). Two characteristics (standardised number of haplotypes/alleles and haplotype diversity/expected heterozygosity) were used to test for concordance between the five loci (cytochrome b and four introns) across all four species of Arripis.
158
Chapter 6. Recent Evolutionary History Kendall’s W was calculated using the online program STATTOOLS (Chang 2009) and the significance assessed using a table of critical values (Siegel and Castellan 1988).
6.2.4.4
Demographic History
Three general approaches were used to examine aspects of the recent demographic history of the four species of Arripis. Firstly, evidence was sought for changes in population size using both mtDNA and nDNA loci. Secondly, where evidence for population expansion was detected in the mtDNA, the date of the start of the expansion was estimated. Finally, estimates of effective population sizes prior to the expansion were made.
6.2.4.4.1
Tests for evidence of changes in population size
6.2.4.4.1.1 Mitochondrial DNA sequences
Several approaches, all of which are based around expectations for mutation-drift equilibrium, were applied to the Arripis cytochrome b sequence data to test for evidence of changes in population sizes (Ramos-Onsins and Rozas 2002). The tests were chosen on the basis that they: (i) are widely used; (ii) are the most appropriate; and (iii) include a representative from each of three classes of test (Ramos-Onsins and Rozas 2002), as described below.
Except where indicated, the statistical significance of each was
assessed by randomly sampling the data 10,000 times in DnaSP version 5.00.04 (Rozas and Rozas 1999), assuming selective neutrality and population equilibrium.
159
Chapter 6. Recent Evolutionary History Class I – The presence of an excess of singleton mutations at segregating sites, indicating recent population expansion, was tested by Tajima’s (1989) D-test and Ramos-Onsins and Ramos’ (2002) R2-test. Both tests were included because the former is based solely on the number of singletons, while the latter compares that number to the mean number of nucleotide differences between sequences. A negative value of D or a low value of R2 would each signify an excess of low frequency polymorphisms, indicating population size expansion and/or directional selection.
Class II – An excess of rare haplotypes relative to that expected based on pairwise nucleotide differences (π) was used to test for population expansion using Fu’s (1997) Fs-test. A negative value of Fs would indicate an excess of rare mutations and hence, population size expansion and/or directional selection.
Class III – Mismatch distributions (Slatkin and Hudson 1991; Rogers and Harpending 1992) of the cytochrome b sequence data were generated for each species and parameters derived from these were used to obtain, by simulation, expected values under a model of: (i) population expansion assuming θ1 = ∞ (i.e. after the growth event); and (ii) stable population size (Rozas and Rozas 1999).
The statistical
significance of any departures between the observed data and the population expansion model was assessed using the sum of squared differences (SSD) and bootstrap approach, as implemented in ARLEQUIN. In this approach, the observed value of SSD was compared to the distribution of SSD values expected for population expansion, generated via 10,000 bootstrap simulations.
The p-value of the SSD statistic was
calculated as the proportion of simulated (expected) SSD values larger than the observed value; a small p-value therefore provides evidence for departure from a model of 160
Chapter 6. Recent Evolutionary History population expansion (Excoffier and Schneider 1999). Another statistic based on the mismatch distribution, raggedness, rg (Harpending 1994) was calculated for each species of Arripis using DnaSP. By providing a measure of the ‘smoothness’ of the mismatch distribution, this test differentiates between unimodal (‘smooth’) curves in expanding populations and multimodal (‘ragged’) curves in stable populations that have accumulated many mutations. Raggedness is an ad hoc test (Harpending 1994), which differs from the SSD method by not testing against a specific model (i.e. expansion). A mismatch distribution could not be generated for the monomorphic A. xylabion.
6.2.4.4.1.2 Nuclear DNA allele frequencies
The intron data (using polymorphic loci only) were used to test for changes in population size with the software BOTTLENECK version 1.2.02 (Cornuet and Luikart 1996) over 10,000 replicate simulations for each locus.
This method used the
theoretical expectation that the number of alleles in a population would be reduced more than the heterozygosity during a population size reduction due to the preferential loss of rare alleles (Nei et al. 1975; Maruyama and Fuerst 1985). The expected heterozygosity of each locus (HE; based on observed allele frequencies, sensu Nei 1987) was compared to a distribution of heterozygosity expected for a population at mutation-drift equilibrium with the observed number of alleles (HExp). This distribution was obtained by coalescent simulation under the infinite-allele model (IAM; Kimura and Crow 1964) and then again for the stepwise mutation model (SMM; Ohta and Kimura 1973). The results were summarised using the equation (HE - HExp), where a positive value suggests a population bottleneck, whereas a negative value suggests a population expansion (Luikart and Cornuet 1998). Although standard nuclear data such as intron length 161
Chapter 6. Recent Evolutionary History polymorphism probably conform more closely to the IAM (Cornuet and Luikart 1996), tests using both IAM and SMM were conducted because they are thought to represent two extremes of mutation models (Chakraborty and Jin 1992), and one major assumption of the IAM is that of non-homology of alleles, and in the present study there is no way to confirm that this was not violated. It was not possible to assess the statistical significance of these results because, as described in Section 6.2.4.2, the recommended approach (Wilcoxon sign–rank test; Cornuet and Luikart 1996) could not be applied to four loci.
6.2.4.4.2
Estimates of timing of population expansions
For each of the species where evidence of population expansion was detected in the mtDNA, the age since the start of the expansion (Harpending 1994) was estimated using two methods which, despite some possible limitations (see Arbogast et al. 2002), are the main approaches used in this context (e.g. Sotka et al. 2005; Janko et al. 2007; Lohman et al. 2008; López et al. 2010). Firstly, the crest of the mismatch distribution curve (see Section 6.2.4.4.1.1), expressed as tau, τ, was estimated with the software DnaSP with θ1 set as infinite (Rogers 1995). Secondly, the coalescent simulation method of Wakeley and Hey (1997) was used to estimate τ in the program SITES (Hey and Wakeley 1997). In both cases, the time since expansion, t, was then calculated by τ = 2ut (Rozas and Rozas 1999; Excoffier et al. 2005), where u is the substitution rate per site per year - in this case 4.83 x 10-6 (1%/site /Myr x length of sequence; Rogers and Harpending 1992, see Section 5.2.2.3.2).
162
Chapter 6. Recent Evolutionary History 6.2.4.4.3
Estimates of effective population sizes
In the species where expansion was detected, the effective female population size Nf (0) at the start of the expansion was estimated. This was estimated using θ = Nf µ, and dividing by the generation time (4.0 years, based on age of sexual maturity for both A. trutta and A. truttaceus; see Chapter 1). Mean values of θ0 (i.e. time = zero) along with 95% confidence estimates, were generated by parametric bootstrap in ARLEQUIN.
6.3
Results
6.3.1
General Information
6.3.1.1
Mitochondrial DNA Sequences
A 483 bp portion of the cytochrome b gene was sequenced for each of 30 individuals of: A. georgianus, 23 A. trutta and 36 A. truttaceus from across their Australian distributions; and 11 A. xylabion from a single locality, and revealed a total of 18 haplotypes (Table 6.1). The average base composition of the sampled portion of the cytochrome b gene was generally similar in all four species of Arripis, although the AT content of the sequence in the A. georgianus samples was slightly less than that in each of the three species of Australian Salmon (Table 6.1). As detailed in Section 4.3.1, there were generally few segregating sites, and substitutions were exclusively synonymous transitional changes at third codon positions in A. truttaceus and A. georgianus, while a single transversion and (rare) first and second codon substitutions were also found in A. trutta (Table 6.1). 163
Chapter 6. Recent Evolutionary History 6.3.1.2
Nuclear DNA Allele Frequencies
Up to 89 individuals of A. trutta, 143 A. truttaceus, 25 A. xylabion, and 261 A. georgianus were assayed for length polymorphism at the four polymorphic intron loci (full details are provided in Table 3.1 and Table 4.2). Locus AldoB4 had the smallest alleles (all < 200 bp) and locus CK7 had the largest (up to 984 bp). The total number of alleles at each locus for all species combined ranged from seven at AldoB4 to 13 at S72 (see Chapter 3; Table 3.2).
Overall polymorphism was greatest in the samples of A. georgianus (four polymorphic loci; total alleles = 32), followed by A. trutta (three polymorphic loci; total alleles = 20), A. truttaceus (two polymorphic loci; total alleles = 9), and A. xylabion (one polymorphic locus; total alleles = 5) (but see standardised results in Section 6.3.2.1). The patterns of variation at the different intron loci were assumed to be independent of each other based on the results of linkage tests (see Chapter 3).
The genotype
frequencies at each locus in each species were generally in accordance with those expected under Hardy Weinberg expectations (see Chapter 3).
164
Table 6.1
Selected characteristics of the 483 bp partial cytochrome b sequences in the species of Arripis. The number of individuals
assayed, the number of haplotypes, the average base composition of the sequence (expressed as percentage), the number of polymorphic sites (i.e. where a nucleotide at a site differs between sequences), the transition/transversion (Ts/Tv) ratio, and the codon positions at which the polymorphic sites were found, are presented. See Table 4.5 for full details of polymorphic sites.
A. georgianus A. trutta A. truttaceus A. xylabion
Number sequenced
Number of haplotypes
%T
%C
%A
%G
30 23 36 11
3 10 4 1
28.8 28.1 28.0 28.0
33.7 32.9 33.1 33.1
23.0 25.3 25.5 25.7
14.5 13.7 13.5 13.3
Number of polymorphic sites 2 10 3 0
Ts/Tv ratio 2/0 9/1 3/0 -
Codon position (1 / 2 / 3) 0/0/2 2/1/7 0/0/3 -
Chapter 6. Recent Evolutionary History 6.3.1.3
Genealogical Relationships
Before presenting the genetic diversity results, it is important to consider the relationships between conspecific cytochrome b haplotypes.
For each of the four
species, the parsimony networks were centred on a common haplotype (Figure 6.2). In A. trutta numerous singletons (haplotypes represented by a single occurrence), which differed from the common haplotype by one or two mutational steps, were present, forming a classic star-shape pattern (Figure 6.2).
This pattern provides strong
circumstantial evidence of the accumulation of new mutations during a population expansion (Avise 2000). Similarly, the parsimony network of A. truttaceus showed a similar star-shape, with a few singletons that were all one mutational step from the common haplotype (Figure 6.2), which is consistent with a very recent population expansion. All sequenced individuals of A. xylabion had identical haplotypes. In A. georgianus, there was one central and two ‘peripheral’ haplotypes, which, although relatively uncommon, were represented by 23% and 7% of individuals (Figure 6.2). Hence, for this species, the network provided no clear evidence of an expansion, although the peripheral haplotypes still differed from the common haplotype by only a single mutational step.
166
Figure 6.2
Parsimony networks for the cytochrome b haplotypes for each species of Arripis. Each line in the network shows a single
mutational change. A haplotype is shown by a circle. The surface area of the circle is proportional to the frequency of each haplotype. Black dots indicate missing haplotypes, which are necessary to link all observed haplotypes present in the network. Sample sizes are indicated in parentheses.
Chapter 6. Recent Evolutionary History 6.3.2
Genetic Diversity
6.3.2.1
General
Both the standardised number of haplotypes and haplotype diversity were noticeably higher in A. trutta (SNH = 10; h = 0.64) than in A. truttaceus (SNH = 3.0; h = 0.16) (Table 6.2). Haplotype diversity was non existent in A. xylabion and moderate in A. georgianus (SNH = 2.9; h = 0.47) (Table 6.2). The level of nucleotide diversity in the partial cytochrome b sequences was very low in each of the four Arripis species (π ≤ 0.002), and especially so in A. truttaceus and A. xylabion (Table 6.2).
Table 6.2
Measures of mtDNA genetic diversity based on the nucleotide
sequence of a 483 bp fragment of the cytochrome b gene in Arripis species. Calculation of standardised number of haplotypes (SNH), haplotype diversity and nucleotide diversity is described in Section 6.2.4.1. Standard errors are provided in parentheses. The monomorphic A. xylabion was not included in the calculations of SNH. Statistic
A. trutta
A. truttaceus
A. xylabion
A. georgianus
n
23
36
11
30
No. haplotypes
10
4
1
3
10.0
3.0
-
2.9
0
(±0.087)
0
(±0.0009)
SNH Haplotype Diversity (h)
0.640
0.162
(±0.014)
(±0.082)
Nucleotide Diversity (π)
0.0020
0.0003
(±0.0020)
(±0.0005)
0.467 0.0010
168
Chapter 6. Recent Evolutionary History The proportion of polymorphic intron loci was low, ranging from 4/12 (PO = 0.33) in A. georgianus to only a single locus in A. xylabion (Table 6.3). The standardised number of alleles across the four polymorphic intron loci ranged from 20 in A. trutta to five in A. xylabion (Table 6.3). Similar standardised values were obtained by a rarefaction approach (Szpiech et al. 2008; results not shown). Average expected heterozygosity ranged from moderate in A. trutta, A. truttaceus and A. georgianus (HE = 0.239 - 0.312) to low in A. xylabion (HE = 0.093) (Table 6.3). Consistent with the mtDNA data, A. trutta exhibited the highest HE, although the standard deviations of the HE estimates were considerable (Table 6.3).
Table 6.3
Measures of nDNA genetic diversity in n individuals of Arripis
species. PO is the proportion of polymorphic loci (of 12), NA is the uncorrected number of alleles, SNA is the standardised number of alleles, and HE is average expected heterozygosity. Standard deviations are provided in parentheses. All estimates except PO were based on only the four loci that were polymorphic in at least one species. Statistic
A. trutta
A. truttaceus
A. xylabion
A. georgianus
n
89
143
25
261
PO
0.25
0.17
0.08
0.33
NA
20
9
5
32
SNA
20
8.6
5
15.2
HE
0.312
0.239
0.093
0.249
(± 0.208)
(± 0.241)
(± 0.161)
(± 0.077)
169
Chapter 6. Recent Evolutionary History 6.3.2.2
Comparisons of Genetic Diversity Between Species
As implied above, the patterns in the mtDNA and nDNA data consistently indicated that A. trutta and A. georgianus had the highest levels of diversity, followed by A. truttaceus and then, A. xylabion, with very low levels of diversity (Tables 6.4, 6.5). An exception was that the number of cytochrome b haplotypes in A. georgianus was less than what might be expected based on the number of intron alleles and intron diversity (Table 6.4; Figure 6.3). This exception held for the number of haplotypes, but not for haplotype diversity (Table 6.5; Figure 6.3).
Table 6.4
Standardised number of haplotypes (cytochrome b) and alleles
(introns) for each locus in the species of Arripis. Species are ranked (parentheses) from highest to lowest for each locus. Mean rank was used for ties (see Section 6.2.4.2). Locus
A. trutta
A. truttaceus
A. xylabion
A. georgianus
cytochrome b
10 (1)
3.1 (2)
1 (4)
3 (3)
AldoB4
6 (1)
2 (3)
1 (4)
4.7 (2)
CaM4
1 (3)
1 (3)
1 (3)
8 (1)
CK7
4 (2)
1 (3.5)
1 (3.5)
6.8 (1)
S72
9 (1)
4.6 (3)
2 (4)
5.2 (2)
170
Chapter 6. Recent Evolutionary History Haplotype diversity (cytochrome b) and expected heterozygosity
Table 6.5
(introns) for each locus in the species of Arripis. Species are ranked (parentheses) from highest to lowest for each locus. Mean rank was used for ties (see Section 6.2.4.2).
Locus
A. trutta
A. truttaceus
A. xylabion
A. georgianus
cytochrome b
0.64 (1)
0.16 (3)
0.00 (4)
0.47 (2)
AldoB4
0.37 (2)
0.44 (1)
0.00 (4)
0.24 (3)
CaM4
0.00 (3)
0.00 (3)
0.00 (3)
0.65 (1)
CK7
0.25 (1)
0.00 (3.5)
0.00 (3.5)
0.23 (2)
S72
0.61 (1)
0.51 (2)
0.36 (4)
0.37 (3)
6.3.3
Neutrality of Markers
Each locus ranked the diversity of the species, relative to each other, in more or less the same order (Tables 6.4 and 6.5) for both haplotype/allele number (W = 0.64, p < 0.01) and haplotype diversity/expected heterozygosity (W = 0.47, p < 0.05). This is consistent with the view that the patterns of diversity at each locus in each species are (more or less) selectively neutral and relate to factors controlling diversity at neutral loci. Therefore, unless otherwise stated, the remainder of the results are interpreted on the assumption that the patterns of variation at the sampled loci are selectively neutral.
171
Chapter 6. Recent Evolutionary History
Standardised N intron alleles
A.
10 8 6 4 2 0 0
1
2
3
4
5
6
7
8
9
10
11
Standardised N cytochrome b haplotypes
Expected heterozygosity
B.
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
cytochrome b haplotype diversity
Figure 6.3
Relationship between cytochrome b and intron locus diversity for each
of: Arripis trutta (yellow), A. truttaceus (pink), A. xylabion (orange) and A. georgianus (blue). A. standardised number of haplotypes against standardised number of alleles. B. haplotype diversity against expected heterozygosity. Intron loci are identified by shape: AldoB4 (♦), CaM4 (■), CK7 (▲), S72 (●).
Graphics have been added to
illustrate where the data points for A. georgianus are outliers (circled) compared to an approximate trend line for the three species of Australian Salmon.
172
Chapter 6. Recent Evolutionary History 6.3.4
Demographic History
Three aspects of demographic history were considered: (i) changes in population size; (ii) the evolutionary age of any such changes; and (iii) the effective population size at the start of such changes. The results of each are presented below.
6.3.4.1
Tests for Evidence of Changes in Population Size
6.3.4.1.1
Mitochondrial DNA sequences
The patterns of cytochrome b sequence variation for A. trutta and, to a lesser degree A. truttaceus, showed evidence of departures from mutation-drift equilibrium and in particular of a population expansion, as follows. Firstly, the values of both Tajima’s D and Fu’s Fs, were negative (and statistically significant; Table 6.6) for the samples of both species, reflecting the excess of rare haplotypes (Section 6.3.1.3). Secondly, the values of R2 for both species were low, as expected under rapid population growth, although bootstrapping for this test indicated a significant disparity from mutation-drift equilibrium only in A. trutta (Table 6.6). Thirdly, the observed mismatch distribution of A. trutta corresponded closely to that predicted under a model of population expansion (Figure 6.4) and did not differ significantly from it (Table 6.6). Although a history of population expansion equally could not be rejected for A. truttaceus (Table 6.6), the extremely low level of polymorphism in this species generated curves that were almost identical for models of either population expansion or a stable population (Figure 6.4). This is typical of species with very low variability and/or those recovering from either a very severe or very recent bottleneck (Harpending 1994; Ramos-Onsins and Rozas 173
Chapter 6. Recent Evolutionary History 2002). Fourthly, a low raggedness index with significant bootstrap support indicated an expanding population in A. trutta, but was not repeated in A. truttaceus (Table 6.6).
The results of the mtDNA-based analyses did not provide strong evidence of a population expansion in A. georgianus although, based on the mismatch SSD, a model of population expansion could not be rejected (Table 6.6; Figure 6.4).
Table 6.6
Tests for evidence of changes in population size in Arripis species.
Test results and associated probabilities (p) are presented for Tajima’s D, Fu’s Fs, Ramos-Onsins and Rozas R2, Harpending’s raggedness rg, and Mismatch SSD.
Statistic
A. trutta
A. truttaceus
A. xylabion
A. georgianus
D
-2.093 p = 0.001
-1.723 p = 0.020
-
-0.024 p = 0.450
Fs
-7.793 p = 0.000
-3.695 p = 0.001
-
0.058 p = 0.501
R2
0.058 p = 0.000
0.092 p = 0.214
-
0.125 p = 0.401
rg
0.042 p = 0.022
0.487 p = 0.593
-
0.173 p = 0.236
SSD
0.0020 p = 0.694
0.0007 p = 0.380
-
0.0540 p = 0.460
174
Chapter 6. Recent Evolutionary History 6.3.4.1.2
Nuclear DNA allele frequencies
The patterns of genetic diversity at the intron loci for A. trutta revealed a heterozygosity deficiency at each polymorphic locus (three), relative to predictions for populations in mutation-drift equilibrium under each mutational model (IAM and SMM), and therefore suggest a population expansion (Table 6.7). In contrast, the patterns for A. truttaceus indicated an excess of heterozygosity at each polymorphic locus (two), for all comparisons except for the S72 locus under the SMM, thus providing some evidence of a relatively recent population bottleneck (Table 6.7).
As explained in Section
6.2.4.4.1.2, it was not possible to test the significance of these differences.
The heterozygosity deficiency at all loci also provides evidence of a population expansion in the Australian Herring A. georgianus (Table 6.7). Little can be drawn from the excess of heterozygosity at the single locus examined in the other species of Australian Salmon A. xylabion (Table 6.7).
175
Chapter 6. Recent Evolutionary History
A.
Frequency
0.6 0.4 0.2 0 0 1 2 3 4 5 6 7 8 9 Number of pairwise differences
Frequency
B.
1 0.5 0 0
1
2
3
4
Number of pairwise differences
C.
Frequency
0.8 0.6 0.4 0.2 0 0
1
2
3
4
5
Number of pairwise differences
Figure 6.4
Mismatch distributions for A. Arripis trutta, B. Arripis truttaceus and
C. Arripis georgianus, indicating the observed frequency of nucleotide differences between pairs of individuals in the sample (columns). The solid line represents the expected distribution under a model of population expansion and the dotted line represents the distribution expected in a stable population.
In A. truttaceus,
expectations for the two models were effectively the same and so both are represented by the single solid line. 176
Chapter 6. Recent Evolutionary History Expected heterozygosity of the actual data (HE) and heterozygosity
Table 6.7
expected at mutation-drift equilibrium given the actual number of alleles under the infinite alleles (IAM), HExpI, and stepwise mutation (SMM), HExpS, models for each polymorphic intron locus in each species of Arripis. Expected values and standard deviation (s.d.) under each model were calculated by 10,000 replications using BOTTLENECK. The prediction of heterozygote deficiency (Hdef) or excess (Hexc) was based on the direction of the equation (HE - HExp).
actual Locus HE Arripis trutta
IAM
SMM
HExpI
s.d.
predict
HExpS
s.d.
predict
AldoB4
0.366
0.532
0.162
Hdef
0.716
0.070
Hdef
CK7
0.254
0.397
0.183
Hdef
0.575
0.110
Hdef
S72
0.606
0.668
0.123
Hdef
0.816
0.040
Hdef
Arripis truttaceus AldoB4
0.442
0.156
0.166
Hexc
0.198
0.170
Hexc
S72
0.514
0.446
0.181
Hexc
0.650
0.090
Hdef
0.227
0.163
Hexc
0.270
0.164
Hexc
Arripis xylabion S72
0.372
Arripis georgianus AldoB4
0.234
0.414
0.191
Hdef
0.641
0.091
Hdef
CaM4
0.158
0.566
0.161
Hdef
0.778
0.052
Hdef
CK7
0.230
0.663
0.131
Hdef
0.842
0.033
Hdef
S72
0.371
0.570
0.159
Hdef
0.780
0.051
Hdef
177
Chapter 6. Recent Evolutionary History 6.3.4.2
Estimates of Timing of Population Expansions
Since strong mtDNA evidence of population expansion was found only in the species of primary interest (A. trutta and A. truttaceus), estimates of the time since the start of the expansion, and the effective female population size at that time, were limited to these two species. Estimates based on both mismatch and coalescent methods indicated that the detected population expansion of A. trutta commenced in excess of 100,000 years ago, but only 17,000 years ago in A. truttaceus, although the confidence intervals for the estimates for each species overlapped (Table 6.8). Associated estimates suggested that the size of the female population at the start of the expansion was higher in A. trutta than in A. truttaceus (Table 6.8).
Table 6.8
Estimates of the time since the start of population expansion and female
effective population size at the start of expansion (Nf
(0))
in Arripis trutta and A.
truttaceus. Estimates are based on mismatch distributions and coalescent simulations. Mutational time, τ, and theta, θ0, were estimated as described in Section 6.2.4.4.2. The 95% confidence intervals derived from bootstrapping are provided for the mismatch method. Method A. trutta
τ
Time (yrs)
θ0
Nf (0)
mismatch
1.028
106,418
0.04
2070
(95% C.I.)
(0.316 - 1.739)
(32,712 – 180,020)
(0 - 0.241)
(0 – 12,474)
1.051
108,799
-
-
coalescent A. truttaceus mismatch
0.167
17,288
0
0
(95% C.I.)
(0.000 - 0.729)
(0 – 75,466)
(0 - 0.005)
(0 - 259)
0.167
17,288
-
-
coalescent
178
Chapter 6. Recent Evolutionary History 6.4
Discussion
All classes of genetic marker studied so far, i.e. the cytochrome b and EPIC markers of this study and the allozyme data of MacDonald (1980) and Ayvazian et al. (2004), indicate that the levels of genetic diversity in all extant representatives of the family Arripidae are relatively low in comparison with the levels typically detected in other abundant marine teleosts (e.g. Grant and Bowen 1998; Hassan et al. 2002). These markers also suggest that the amount of diversity varies more or less predictably among different members of the Arripidae, with species in order from highest diversity to lowest being A. trutta, A. georgianus, A. truttaceus and A. xylabion (notwithstanding the limited spatial sampling that was conducted for this last species). The above two patterns of diversity are likely to be linked strongly to demographic stochasticity in these species, all of which may be particularly susceptible to the effects of glacial cycles (see below). Overall the mtDNA and nDNA data indicate that the populations of Australian Salmon east and west of the Bass Strait are not in mutation-drift equilibrium and suggest that A. truttaceus has undergone a more recent and/or more significant population expansion compared to A. trutta. This latter finding is consistent with the a priori prediction that A. truttaceus has been more severely impacted by glacial cycles than A. trutta by virtue of its distribution to the west of Bass Strait (see Section 6.1.1).
6.4.1
Genetic Diversity in Arripis
While it is difficult to compare across studies, the available evidence indicates that the levels of genetic diversity in the extant species of the Arripidae are relatively low compared to those typically seen in other abundant marine teleosts. This is despite the 179
Chapter 6. Recent Evolutionary History fact that data for each of three of the four arripid species are based on samples from across the entire Australian distribution of the species. For example, in data reported for the cytochrome b, ND4/5 or cytochrome oxidase regions of the mtDNA for 31 marine teleosts in Table 4 of Grant and Bowen (1998), 23 (74%) and 22 (71%) species had values of haplotype diversity and nucleotide diversity, respectively, that were higher than those of A. trutta (the most diverse of the Arripidae in this respect). In addition, in data reported for intron length polymorphism for 13 marine teleosts in Hassan et al. (2002), 9 (69%) species had a higher proportion of polymorphic loci than A. georgianus (the most diverse of the Arripidae in this respect). Furthermore, Parker et al. (1998) reported that the average proportion of polymorphic allozyme loci in 183 species of fishes was 21%, which is considerably higher than the highest value recorded in a species of Arripis (A. trutta with 3/27 (11%); MacDonald 1980). The fact that the results for each Arripis species were fairly consistent over all classes of marker studied so far suggests that the low levels of diversity in each species are genome-wide. However, more intensive sampling of the genome, e.g. via techniques, such as amplified fragment length polymorphism, that simultaneously amplify a large number of loci (e.g. Fischer et al. 2011), is required to confirm this.
The exact reason for the generally low levels of genetic diversity in the Australian arripids is not clear, but in general terms must reflect either a low rate of molecular evolution and/or a relatively high rate of loss of genetic variation, which could be driven by either genetic drift and/or natural selection, in each of these species (see Amos and Harwood 1998). The rate of molecular evolution in the family Arripidae is not known and cannot readily be determined (see Section 5.2.2.3.2), however there is no compelling reason to believe that this rate is dramatically different to that in other 180
Chapter 6. Recent Evolutionary History teleosts. Genome-wide reductions in diversity are often attributed to a small effective population size associated with either a persistently small population size and/or a population bottleneck(s) (e.g. Watterson 1984; Avise 2000; Campos et al. 2010), although Amos and Harwood (1998) have argued that this explanation is rarely sufficient on its own and that selective sweeps may be more important than generally recognised.
Regardless, a small effective population size provides a plausible
explanation for the low levels of genetic diversity in each Arripis species, particularly if one invokes recurrent bottlenecks associated with Pleistocene glacial events and their effects on coastal environments.
If the low levels of genetic diversity in each Arripis species are mainly due to Pleistocene bottlenecks, then this implies that these species are more susceptible to the effects of glaciation than many other teleosts that occur in temperate coastal waters (because many such species do not show such low diversity). Although highly mobile, the Arripis species could nevertheless be particularly susceptible to the effects of glacial events if, for example, current flows that maintain critical linkages between larval, juvenile and adult distributions (see Gillanders et al. 2003) are disrupted (see Chapter 1 for the details of the unusually complex life-cycle of these species). It is also possible that the Arripis species have particularly narrow thermal tolerances and/or range of spawning temperature. While there is no specific evidence to support or refute this idea, it is relevant that, where known, spawning in each species is highly temporally and spatially restricted (see Chapter 1) and temperature is likely to be an important cue in the formation of the spawning aggregations (Stewart et al. 2011).
181
Chapter 6. Recent Evolutionary History 6.4.2
Recent Evolutionary History of Arripis trutta and Arripis truttaceus
Both the mtDNA data (e.g. ‘star’ phylogenies, neutrality tests, mismatch distributions) and the EPIC data (e.g. heterozygosity deficiency) provide evidence of population expansion in A. trutta, with the expansion reflected in the mtDNA dated to around 100,000 years ago. In contrast, while the mtDNA data also suggest a population expansion in A. truttaceus (as per A. trutta), but starting considerably more recently than A. trutta (around 20,000 years ago), the intron loci (e.g. heterozygosity excess) suggest a contraction.
The two different classes of marker suggest different demographic
scenarios (expansion and contraction) for A. truttaceus probably because they are each measuring demographic changes over different time-scales due to different rates of mutation (e.g. Eytan and Hellberg 2010). In particular, since the mutation rate of a vertebrate mtDNA locus is typically around ten times faster than that of a nDNA locus (e.g. Brown et al. 1979, but see Ballard and Whitlock 2004), it is likely that A. truttaceus experienced an expansion following a contraction. In the case of A. trutta, assuming that this species also underwent a contraction before the detected expansion(s), evidence of the contraction may no longer be apparent in either the EPIC or mtDNA data if the contraction occurred a relatively long time ago.
The results of this chapter indicate that A. truttaceus has experienced more dramatic and/or recent population expansion than A. trutta, as outlined below.
Firstly, the levels of genetic diversity are consistently higher in A. trutta compared to A. truttaceus, although the difference could not be tested statistically based on the data from the four polymorphic loci of this study. However, by also including data for five 182
Chapter 6. Recent Evolutionary History polymorphic allozyme loci from MacDonald (1980), a Wilcoxon sign-rank test (Gehring 1978) demonstrated that A. trutta had a significantly higher diversity than A. truttaceus for both the standardised number of haplotypes/alleles (p = 0.004) and haplotype diversity/expected heterozygosity (p = 0.019) (data not shown).
Secondly, while the mismatch distributions of the mtDNA sequence data (and associated neutrality tests) provide evidence of population expansions in both A. trutta and A. truttaceus, the estimated time of the start of the expansion in the latter species coincides with the end of the peak LGM, some 20,000 years ago (see Lambeck et al. 2002), which is considerably more recent than that estimated for the former (~ 100,000 years ago).
Thirdly, the results of Cornuet and Luikart’s (1996) bottleneck tests using intron loci imply a population expansion in A. trutta and a population bottleneck in A. truttaceus, although they are based on only a small number of loci with no statistical testing. Note that, although a heterozygosity deficit was suggested at one nDNA locus in A. truttaceus under the SMM, Cornuet and Luikart (1996) have shown that, under this model, some loci may show such a deficiency even during a bottleneck. Furthermore, the SMM is the model considered least likely to apply to the intron length frequency data (see Section 6.2.4.4.1.2). Adding the allozyme data (of MacDonald 1980, as above) to the bottleneck tests was considered, but there were few polymorphic loci, with very few alleles and extremely low HE, rendering their addition to the analyses effectively meaningless or worse, erroneous (see Cornuet and Luikart 1996).
183
Chapter 6. Recent Evolutionary History The apparent differences in the demographic histories of A. trutta and A. truttaceus are probably related to distributional differences between the species (e.g. Colson and Hughes 2007; Janko et al. 2007; McGaughran et al. 2010) rather than to biological or ecological differences (e.g. Fauvelot et al. 2003; Hickerson and Cunningham 2005) This is because the ecological and biological differences between these two species are in fact minimal (see Chapter 1). On the other hand, the results are consistent with expected differences in the impacts of glacial events on A. trutta (lesser) and A. truttaceus (greater) by virtue of their respective distributions to the east and west of the Bassian Isthmus (see Section 6.1.1).
The above conclusion that A. truttaceus was more severely impacted by the LGM (and/or other glacial events) than A. trutta is dependent on several assumptions. This includes the assumption that the rate of mutation is similar in both species (rather than higher in A. trutta), which seems reasonable given that they are closely related species (see Avise 2000; Arbogast et al. 2002). The conclusion is also partly based on the results of the bottleneck tests, which ideally should include a much larger number of loci (Cornuet and Luikart 1996). Similarly, the dating of the demographic changes is dependent upon accurate knowledge of the rate of molecular evolution in the Arripidae (which is not known, see Section 5.2.2.3.2), although it is noteworthy that the estimate of the start of the expansion in A. truttaceus coincides with the peak of the LGM. Furthermore, this conclusion also assumes that the results reflect aspects of the demographic histories of each species rather than a differential capture of polymorphism from their most recent common ancestor, and is supported by the estimates that date the demographic changes to more than 500,000 years after speciation (see Chapter 5). Finally, linking demographic stochasticity to Pleistocene glacial events 184
Chapter 6. Recent Evolutionary History is necessarily based on inferential evidence (e.g. Avise 2000; Bowen et al. 2001; Beheregaray et al. 2002; Bernardi 2005; Crandall et al. 2008; López et al. 2010), although one of the strengths of this study is that it used sibling species to test a priori predictions about the impacts of glacial events. Despite these assumptions, multiple aspects of the data set are nevertheless consistent with the above conclusion.
This is the second study to provide genetic evidence in support of the view that populations to the west of the Bass Strait have been more severely impacted by glacial events than comparable populations to the east. The first study compared eastern and western populations of the Bull Kelp Durvillaea potatorum (Fraser et al. 2009). In addition, other single-species studies have found higher levels of genetic diversity in eastern populations compared to their western counterparts in a gastropod (Brown 1991) and a fish (Colgan and Paxton 1997), although these studies were descriptive rather than providing explicit tests regarding the differential impact of glaciation events on the east and west side of the Bass Strait. The present study is unique in that it has compared eastern and western sibling species and has therefore potentially provided a longer-term perspective.
6.4.3
Comparison of Arripis truttaceus and Arripis georgianus
In contrast to A. truttaceus, with which it is broadly co-distributed, the Australian Herring A. georgianus did not show clear evidence of a population contraction followed by an expansion linked to the LGM. This implies that the impact of the LGM was not as great on A. georgianus as for A. truttaceus. Different responses to the LGM in these two species may be linked to subtle, but important, differences in their distributions. In 185
Chapter 6. Recent Evolutionary History particular, the present distribution of A. georgianus extends much further up the west coast of Australia than that of A. truttaceus (see Chapter 1) and the temperature effects of the LGM were probably less severe along the west coast than on the south coast where the latter species mainly resides (see Section 6.1.1; Figure 6.1). In addition, A. georgianus may differ in some biological attributes (see Chapter 1) that enhance resilience to glacial changes (see Kaustuv et al. 2001; Wares and Cunningham 2001; Rockwood 2006). For example, the diet of A. georgianus is much more general than that of A. truttaceus and resilience to glacial changes is thought to be higher among dietgeneralists (Hickerson and Cunningham 2005).
6.4.4
Recent Evolutionary History of Arripis xylabion
The sample of the third species of Australian Salmon, A. xylabion from Lord Howe Island had almost no genetic diversity, although it is not clear if more genetic diversity is present elsewhere in the species distribution. This lack of diversity prohibited a detailed analysis of the recent evolutionary history of this species. The low diversity at Lord Howe Island may reflect a: (i) persistently small population size; and/or (ii) severe bottleneck (or series of bottlenecks); and/or (iii) founding of the population by a small number of individuals.
Although it is not possible to distinguish among these
alternatives, it has been proposed that this species arose via peripheral isolation (founder effect) in the mid to late Pleistocene (Chapter 5) and thus may have started with relatively low levels of genetic diversity. It is also likely that the population of this species at Lord Howe Island may have experienced repeated bottlenecks in response to subsequent glacial events and/or other factors (see Bostock et al. 2006; Woodroffe et al. 2006). For example, the fringing coral reefs of the island are Holocene in origin 186
Chapter 6. Recent Evolutionary History implying that the fauna was influenced by the LGM and the subsequent Holocene marine transgression (see Woodroffe et al. 2006).
6.4.5
Conclusions
This study provides one of the first direct tests of the prediction that the populations to the west of the Bass Strait may have been more severely impacted by glacial cycles, relative to their east coast counterparts.
The evidence was consistent with this
prediction. In particular, the evidence suggests that A. truttaceus, which resides to the west of the Bass Strait, may have experienced a severe population bottleneck during the LGM followed by an expansion commencing some 20,000 years ago. In contrast, the population of A. trutta, which resides to the east of the Bass Strait, appears to have been largely unaffected by the LGM and has been expanding over the past 100,000 years or more. This general finding needs to be confirmed by examining other species, and the results for A. georgianus illustrate the value of conducting such comparisons using comparable populations (conspecific or sibling species) in order to minimise the effect of confounding factors.
187
Chapter 7. General Conclusions
7.0
General Conclusions
The purpose of this final chapter is to summarise the main conclusions of each of the data chapters. This study used the patterns of variation in a 483 bp fragment of the cytochrome b gene and a 654 bp fragment of the COI gene in the mtDNA, and at four nuclear intron loci to investigate aspects of the population genetic structure, phylogenetic relationships and recent evolutionary history of four species of Arripis. The results of this study indicate the following:
1.
Fragment length polymorphism at EPIC amplified intron loci is generally an
effective approach for generating nuclear data using universal primers and avoiding the high costs associated with the development of other nDNA markers (e.g. microsatellites). Despite the use of capillary electrophoresis, binning of alleles into size classes was required in order to account for scoring variabilities and this may have masked some genetic variation or underestimated some measures of genetic diversity. This study found that EPIC primers that amplify shorter fragments (< 500 bp) should be favoured because they give highly consistent results, are polymorphic and the resultant data require considerably less binning. It also worth noting that the advent of next generation sequencing, which was not readily available during the planning phase of this research, has made the development of novel microsatellite markers more affordable (Santana et al. 2009; Metzker 2010).
2.
There was no evidence for spatial genetic heterogeneity across almost the entire
distributions of any of the three species of Arripis in Australian waters – A. trutta, A. truttaceus and A. georgianus. 188
Chapter 7. General Conclusions 3.
On the basis of knowledge of their life history, each species of Arripis probably
has considerable gene flow across its distribution and each might comprise a single, panmictic population in Australian waters, with genetic homogeneity promoted by: (i) an extensive pelagic planktonic larval stage; (ii) extensive mixing by migrating juveniles and sub-adults; (iii) highly vagile and dispersive adults; and (iv) a relatively restricted spawning area.
4.
The current four-species taxonomy of Arripis based on morphological data (i.e.
Paulin 1993), was supported by molecular data on the basis that there were fixed nucleotide differences in mtDNA sequences (at least in two species) accompanied by significant differentiation in nDNA allele frequencies, interspecific genetic distances that were substantially greater than intraspecific distances and reciprocal monophyly inferred by all phylogenetic methods.
5.
There was no definitive phylogenetic support for the putative geminate status of
the species east and west of the Bass Strait (i.e. A. trutta and A. truttaceus, respectively). There was considerable evidence for a trifurcating polytomy where the three species of Australian Salmon (A. trutta, A. truttaceus, A. xylabion) arose at, or nearly at, the same time, although inferences may be limited by the resolving power of the data.
6.
Regardless of the relationship between the Australian Salmon, the three species
were estimated to have diverged from the most recent common ancestor in the mid to late Pleistocene, around 650,000 to 700,000 years ago.
189
Chapter 7. General Conclusions 7.
Relative to other marine fishes, all four species of Arripis appear to exhibit
generally low levels of genetic diversity in both the mtDNA and at nuclear intron loci. This is consistent with the low diversity reported for allozyme markers by MacDonald (1980).
8.
Among the three closely related Australian Salmon species there was a
consistent trend across all loci where A. trutta showed the highest genetic diversity, A. truttaceus had very low levels of diversity and A. xylabion almost none. The Australian Herring A. georgianus also showed a high diversity, relative to other Arripis, in all measures except the number of cytochrome b haplotypes.
9.
The species of Australian Salmon to the west of the Bass Strait, A. truttaceus,
may have been more severely impacted by glacial cycles, relative to A. trutta on the east coast.
The evidence suggests that A. truttaceus may have experienced a severe
population bottleneck during the Last Glacial Maximum followed by an expansion commencing some 20,000 years ago. In contrast, by virtue of its distribution, the population of A. trutta appears to have been largely unaffected by the LGM and has been expanding over the past 100,000 years or more.
10.
A concordant response to the Last Glacial Maximum was not detected between
the two westerly distributed species, A. truttaceus and A. georgianus. The genetic patterns in A. georgianus suggest that this species may have survived the LGM better than its sympatric congener A. truttaceus due to a combination of more westerly distribution and possible biological differences.
190
Chapter 7. General Conclusions 11.
In summary, this study finds evidence of different evolutionary histories for
species with east/west allopatric distributions across southern Australia. There are many putative geminate species in the region and this study should encourage many more similar studies. As suggested by Burridge (2000b), such studies should examine genetic variation in taxa across different distributional ranges, habitat requirements, life histories and dispersal capabilities. This study also adds weight to the importance of the Bass Strait as an example of a significant temperate biogeographical barrier and its role in shaping the temperate marine fauna of the Australian region. Given that the Bass Strait is now well established as a genetic isolating barrier (e.g. Burridge 1999;2000b; Waters and Roy 2003; Waters et al. 2004; Dawson 2005; York et al. 2008; this study), future studies could focus on addressing the paucity of genetic data comparing specieslevel population and recent evolutionary histories. The many putative geminate species in the region provide an excellent opportunity for comparative studies in this context, especially for explicit hypothesis testing.
191
References
References Almuly, R., B. Cavari, H. Ferstman, O. Kolodny and B. Funkenstein (2000). Genomic structure and sequence of the gilthead seabream (Sparus aurata) growth hormoneencoding gene: Identification of minisatellite polymorphism in intron 1. Genome 43: 836-845. Als, T.D., M.M. Hansen, G.E. Maes, M. Castonguay, L. Riemann, K. Aarestrup, P. Munk, H. Sparholt, R. Hanel and L. Bernatchez (2011). All roads lead to home: panmixia of European eel in the Sargasso Sea. Molecular Ecology 20(7): 13331346. Amos, W. and J. Harwood (1998). Factors affecting levels of genetic diversity in natural populations. Philosophical Transactions of the Royal Society (London) Series B 353: 177-186. Andre´, C., L.C. Larsson, L. Laikre, D. Bekkevold, J. Brigham, G.R. Carvalho, T.G. Dahlgren, W.F. Hutchinson, S. Mariani, K. Mudde, D.E. Ruzzante and N. Ryman (2011). Detecting population structure in a high gene-flow species, Atlantic herring (Clupea harengus): direct, simultaneous evaluation of neutral vs putatively selected loci. Heredity 106: 270-280. Anon. (2004). Microsatellite Analysis on the Applied Biosystems 3130 Series Genetic Analyzers. from www.appliedbiosystems.com. Anon. (2010). Kahawai, New Zealand Fisheries. Arbogast, B.S., S.V. Edwards, J. Wakeley, P. Beerli and J.B. Slowinski (2002). Estimating divergence times from molecular data on phylogenetic and population genetic timescales. Annual Review of Ecology and Systematics 33: 707-740. Atarhouch, T., M. Rami, G. Cattaneo-Berrebi, C. Ibanez, S. Augros, E. Boissin, A. Dakkak and P. Berrebi (2003). Primers for EPIC amplification of intron sequences for fish and other vertebrate population genetic studies. BioTechniques 35(4): 676682. Atarhouch, T., M. Rami, M. Naciri and A. Dakkak (2007). Genetic population structure of sardine (Sardina pilchardus) off Morocco detected with intron polymorphism (EPIC-PCR). Marine Biology 150: 521-528. Avise, J.C. (1986). Mitochondrial DNA and the evolutionary genetics of higher animals. Philosophical Transactions of the Royal Society (London) Series B 312: 325-342. Avise, J.C. (2000). Phylogeography. The History and Formation of Species. Cambridge, Massachusetts, Harvard University Press. Avise, J.C. (2004). Molecular Markers, Natural History, and Evolution. Sunderland, Massachusetts, Sinauer Associates, Inc. Avise, J.C. (2009). Phylogeography: retrospect and prospect. Journal of Biogeography 36: 3-15. Avise, J.C., J. Arnold, R.M. Ball, E. Bermingham, T. Lamb, J.E. Neigel, C.A. Reeb and N.C. Saunders (1987). Intraspecific Phylogeography: The mitochondrial DNA bridge between population genetics and systematics. Annual Review of Ecology and Systematics 18: 489-522. Avise, J.C. and R.M. Ball (1990). Principles of genealogical concordance in species concepts and biological taxonomy. Oxford Surveys in Evolutionary Biology 7: 4567. Ayre, D.J. and T.P. Hughes (2004). Climate change, genotypic diversity and gene flow in reef-building corals. Ecology Letters 7: 273-278. 192
References Ayre, D.J., T.E. Minchinton and C. Perrin (2009). Does life history predict past and current connectivity for rocky intertidal invertebrates across a marine biogeographic barrier? Molecular Ecology 18: 1887-1903. Ayvazian, S.G., T.P. Bastow, J.S. Edmonds, J. How and G.B. Nowara (2004). Stock structure of Australian Herring (Arripis georgiana) in southwestern Australia. Fisheries Research 67: 39-53. Baack, E.J. and L.H. Rieseberg (2007). A genomic view of introgression and hybrid speciation. Current Opinions in Genetics and Development 17: 513-518. Baker, A.N. (1971). Food and feeding of Kahawai (Teleostei: Arripididae). New Zealand Journal of Marine and Freshwater Research 5(2): 291-299. Ballard, J.W.O., M.K. Thayer, A.F. Newton and E.R. Grismer (1998). Data sets, partitions, and characters: philosophies and procedures for analyzing multiple data sets. Systematic Biology 47: 367-396. Ballard, J.W.O. and M.C. Whitlock (2004). The incomplete natural history of mitochondria. Molecular Ecology 13: 729-744. Bandelt, H.-J. and A.W.M. Dress (1992). Split decomposition: A new and useful approach to phylogenetic analysis of distance data. Molecular Phylogenetics and Evolution 1: 242-252. Barth, D., E. Przyboś, S.I. Fokin, M. Schlegel and T.U. Berendonk (2008). Cytochrome b sequence data suggest rapid speciation within the Paramecium aurelia species complex. Molecular Phylogenetics and Evolution 49(2): 669-673. Barton, M. (2007). Bond's biology of fishes. Belmont, California, Thomson. Baum, D.A. and K.L. Shaw (1995). Genealogical perspectives on the species problem. Experimental and Molecular Approaches to Plant Biosystematics. P.C. Hoch and A.G. Stephenson. St. Louis, MO., Missouri Botanical Garden. Bearham, D. (2004). A mtDNA study of the population structure of silver trevally Pseudocaranx dentex and the relationship between silver and sand trevally Pseudocaranx wrighti in Western Australian waters. School of Biological Sciences. Perth, Western Australia, Murdoch University: 112. Beaumont, M.A., R. Nielsen, C. Robert, J. Hey, O. Gaggiotti, L.L. Knowles, A. Estoup, M. Panchal, J. Corander, M.J. Hickerson, S.A. Sisson, N. Fagundes, L. Chikhi, P. Beerli, R. Vitalis, J.M. Cornuet, J.P. Huelsenbeck, M. Foll, Z. Yang, F. Rousset, D. Balding and L. Excoffier (2010). In defence of model-based inference in phylogeography. Molecular Ecology 19: 436–446. Beheregaray, L.B. (2008). Twenty years of phylogeography: the state of the field and the challenges for the Southern Hemisphere. Molecular Ecology 17: 3754-3774. Beheregaray, L.B., P. Sunnucks and D.A. Briscoe (2002). A rapid radiation associated with the last sea-level changes in southern Brazil: the silverside Odontesthes perugiae complex. Proceedings of the Royal Society B 269: 65-73. Belkhir, K., P. Borsa, J. Goudet, L. Chikhi and F. Bonhomme (2004). GENETIX, logiciel sous Windows™ pour la génétiques des populations. Montpellier, France, Laboratoire Génome et Populations, Université Montpellier II. Belyaeva, N.V. and I.I. Burmistrova (1992). Some features of the Pleistocene paleocirculation in the eastern equatorial Indian Ocean (by Foraminifera). Centenary of Japanese Micropaleontology. K. Ishizaki and T. Saito. Tokyo, Terra Scientific Publishing Company: pp. 67-70. Bennett, I. and E.C. Pope (1953). Intertidal zonation of the exposed rocky shores of Victoria, together with a rearrangement of the biogeographical provinces of 193
References temperate Australian shores. Australian Journal of Marine and Freshwater Research 4: 105-159. Benson, G.A., I. Karsch-Mizrachi, D.J. Lipman, J. Ostell and D.L. Wheeler (2007). GenBank, National Center for Biotechnology Information, National Library of Medicine, Bethesda, USA. Bermingham, E., S.S. McCafferty and A.P. Martin (1997). Fish biogeography and molecular clocks: Perspectives from the Panamanian Isthmus. Molecular Systematics of Fishes. T.D. Kocher and C.A. Stepien. San Diego, Academic Press: pp. 113-128. Bernardi, G. (2005). Phylogeography and demography of sympatric sister surfperch species Embiotoca jacksoni and E. lateralis along the Californian coast: historical versus ecological factors. Evolution 59: 386-394. Bernardi, G., L. Findley and A. Rocha-Olivares (2003). Vicariance and dispersal across Baja California in disjunct marine fish populations. Evolution 57(7): 1599-1609. Bernardi, G. and J. Lape (2005). Tempo and mode of speciation in the Baja California disjunct fish species Anisotremus davidsonii. Molecular Ecology 14: 4085-4096. Berrebi, P., E. Boissin, F. Fang and G. Cattaneo-Berrebi (2005). Intron polymorphism (EPIC-PCR) reveals phylogeographic structure of Zacco platypus in China: a possible target for aquaculture development. Heredity 94: 589-598. Berrebi, P., X. Retif, F. Fang and C. Zhang (2006). Population structure and systematics of Opsariichthys bidens (Osteichthys: Cyrprinidae) in south-east China using a new nuclear marker: the introns (EPIC-PCR). Biological Journal of the Linnean Society 87: 155-166. Bierne, N., S.A. Lehnert, E. Bédier, F. Bonhomme and S.S. Moore (2000). Screening for intron-length polymorphisms in penaeid shrimps using exon-primed introncrossing (EPIC)-PCR. Molecular Ecology 9: 233-235. Birky, J.C.W. (2001). The inheritance of genes in mitochondria and chloroplasts: laws, mechanisms, and models. Annual Review of Genetics 35: 125-148. Bloomquist, E.W., P. Lemey and M.A. Suchard (2010). Three roads diverged? Routes to phylogeographic inference. Trends in Ecology and Evolution 25(11): 626-632. Bolton, T.F., C.J. Hayward and A.J. Turner (2006). The Piscicolid leech Austrobdella translucens Badham 1916: a new host record from Australian salmon, Arripis trutta (Forster 1801), and a new locality record for Yellowfin whiting, Sillago schomburgkii Peters 1864, in South Australia. Australian Zoologist 33(3): 385-387. Bonin, A., E. Bellemain, P. Bronken Eidesen, F. Pompanon, C. Brochmann and P. Taberlet (2004). How to track and assess genotyping errors in population genetics studies. Molecular Ecology 13: 3261–3273. Bostock, H.C., B.N. Opdyke, M.K. Gagan, A.E. Kiss and L.K. Fifield (2006). Glacial/interglacial changes in the East Australian Current. Climate Dynamics 26: 645-659. Bowen, B.W. and J.C. Avise (1990). Genetic structure of Atlantic and Gulf of Mexico populations of sea bass, menhaden, and sturgeon: Influence of zoogeographic factors and life-history patterns. Marine Biology 107: 371-381. Bowen, B.W., A.L. Bass, L.A. Rocha, W.S. Grant and D.R. Robertson (2001). Phylogeography of the trumpetfishes (Aulostomus): ring species complex on a global scale. Evolution 55: 1029–1039. Boyle, M.C., N.N. FitzSimmons, C.J. Limpus, S. Kelez, X. Velez-Zuazo and M. Waycott (2009). Evidence for transoceanic migrations by loggerhead sea turtles in the southern Pacific Ocean. Proceedings of the Royal Society B 276: 1993-1999. 194
References Braun, E.L. and R.T. Kimball (2001). Polytomies, the power of phylogenetic inference, and the stochastic nature of molecular evolution: a comment on Walsh et al. (1999). Evolution 55: 1261-1263. Brito, P. and A.W. Edwards (2008). Multilocus phylogeography and phylogenetics using sequence based markers. Genetica 135: 439-455. Bromham, L. and D. Penny (2003). The modern molecular clock. Nature Reviews Genetics 4: 216-224. Bromham, L., M. Woolfit, M.S.Y. Lee and A. Ranmbaut (2002). Testing the relationship between morphological and molecular rates of change along phylogenies. Evolution 56(10): 1921-1930. Broquet, T. and E. Petit (2003). Quantifying genotyping errors in noninvasive population genetics. Molecular Ecology 13: 3601–3608. Broughton, R.E. and T.E. Dowling (1997). Evolutionary dynamics of tandem repeats in the mitochondrial DNA control region of the Minnow Cyprinella spiloptera. Molecular Biology and Evolution 14(12): 1187-1196. Brown, L.D. (1991). Genetic variation and population structure in the blacklip abalone, Haliotis rubra. Australian Journal of Marine and Freshwater Research 42: 77-90. Brown, W.M., M. George and A.C. Wilson (1979). Rapid evolution of animal mitochondrial DNA. Proceedings of the National Academy of Sciences 76: 19671971. Bruce, B.D., F.J. Neira and A.G. Miskiewicz (1998). Arripidae: Australian salmons. Larvae of Temperate Australian Fishes. F.J. Neira, A.G. Miskiewicz and T. Trnski. Nedlands, University of Western Australia Press: 180-182. Bryant, D. and V. Moulton (2004). NeighborNet: an agglomerative method for the construction of phylogenetic networks. Molecular Biology and Evolution 21: 255265. Burridge, C.P. (1999). Molecular Phylogeny of Nemadactylus and Acantholatris (Perciformes: Cirrhitoidea: Cheilodactylidae), with implications for taxonomy and biogeography. Molecular Phylogenetics and Evolution 13(1): 93-109. Burridge, C.P. (2000a). Molecular phylogeny of the Aplodactylidae (Perciformes: Cirrhitiodea), a group of Southern Hemisphere marine fishes. Journal of Natural History 34: 2173-2185. Burridge, C.P. (2000b). Biogeographic history of geminate cirrhitoids (Perciformes: Cirrhitoidea) with east-west allopatric distributions across southern Australia, based on molecular data. Global Ecology and Biogeography 9: 517-525. Burridge, C.P. (2002). Antitropicality of Pacific fishes: molecular insights. Environmental Biology of Fishes 65: 151-164. Burridge, C.P., R.C. Meléndez and B.S. Dyer (2006). Multiple origins of the Juan Fernández Kelpfish fauna and evidence for frequent and unidirectional dispersal of cirrhitoid fishes across the South Pacific. Systematic Biology 55(4): 566-578. Burridge, C.P. and R.W.G. White (2000). Molecular phylogeny of the antitropical subgenus Goniistius (Perciformes: Cheilodactylidae: Cheilodactylus): evidence for multiple transequatorial divergences and non-monophyly. Biological Journal of the Linnean Society 70: 435-458. Cambray, J.A. (1994). The comparative reproductive styles of two closely related African minnows (Pseudobarbus afer and P. asper) inhabiting two different sections of the Gamtoos River system. Environmental Biology of Fishes 41: 247268. 195
References Campos, P.F., T. Kristensen, L. Orlando, A. Sher, M.V. Kholodova, A. Gotherstrom, M. Hofreiter, D.G. Drucker, P. Kosintsev, A. Tikhonov, G.F. Baryshnikov, E. Willerslev and M.T.P. Gilbert (2010). Ancient DNA sequences point to a large loss of mitochondrial genetic diversity in the saiga antelope (Saiga tatarica) since the Pleistocene. Molecular Ecology 19(22): 4863-4875. Cappo, M., C.J. Walters and R.J. Lenanton (2000). Estimation of rates of migration, exploitation and survival using tag recovery data for western Australian "salmon" (Arripis truttaceus: Arripidae: Percoidei). Fisheries Research 44: 207-217. Caputi, N., W.J. Fletcher, A. Pearce and C.F. Chubb (1996). Effect of the Leeuwin Current on the recruitment of fish and invertebrates along the Western Australian coast. Marine and Freshwater Research 47: 147-155. Cárdenas, L., A.X. Silva, A. Magoulas, J. Cabezas, E. Poulin and F.P. Ojeda (2009). Genetic population structure in the Chilean jack mackerel, Trachurus murphyi (Nichols) across the South-eastern Pacific Ocean. Fisheries Research 100: 109115. Catalano, S.R. and K.S. Hutson (2010). Harmful parasitic crustaceans infecting wild arripids: a potential threat to southern Australian finfish aquaculture. Aquaculture 303: 101-104. Catalano, S.R., K.S. Hutson, R.M. Ratcliff and I.D. Whittington (2010). Redescriptions of two species of microcotylid monogeneans from three arripid hosts in southern Australian waters. Systematic Parasitology 76: 211-222. Catalano, S.R., K.S. Hutson, R.M. Ratcliff and I.D. Whittington (2011). The value of host and parasite identification for arripid fish. Marine and Freshwater Research 62(1): 72-82. Chairi, H., M. Idaomar and L. Rebordinos (2007). Mitochondrial DNA analysis of the European Anchovy in the southern Mediterranean and northern Atlantic coasts. Journal of Fisheries and Aquatic Science 2(3): 206-215. Chakraborty, R. and L. Jin (1992). Heterozygote deficiency, population substructure and their implications in DNA fingerprinting. Human Genetics 88: 267-272. Chang, A.M. (2009). STAT TOOLS. Chen, I.-S., C.-H. Hsu, C.-F. Hui, K.-T. Shao, P.J. Miller and L.-S. Fang (1998). Sequence length and variation in the mitochondrial control region of two freshwater gobiid fishes belonging to Rhinogobius (Teleostei: Gobioidei). Journal of Fish Biology 53: 179-191. Chenoweth, S.F., J.M. Hughes, C.P. Keenan and S. Lavery (1998). When oceans meet: a teleost shows secondary intergradation at an Indian-Pacific interface. Proceedings of the Royal Society B 265(1394): 415-420. Chlaida, M., V. Laurent, S. Kifani, T. Benazzou, H. Jaziri and S. Planes (2008). Evidence of a genetic cline for Sardina pilchardus along the Northwest African coast. ICES Journal of Marine Science 66: 264-271. Chow, S. and K. Hazama (1998). Universal PCR primers for S7 ribosomal protein gene introns in fish. Molecular Ecology 7: 1247-1263. Chow, S. and M. Nakadate (2004). PCR primers for fish G6PD gene intron and characterization of intron length variation in the albacore Thunnus alalunga. Molecular Ecology Notes 4: 391-393. Chow, S., V.P. Scholey, A. Nakazawa, D. Margulies, J.B. Wexler, R.J. Olson and K. Hazama (2001). Direct evidence for mendelian inheritance of the variations in the ribosomal protein gene introns in Yellowfin Tuna (Thunnus albacares). Marine Biotechnology 3: 22-26. 196
References Chow, S. and H. Takeyama (2000). Nuclear and mitochondrial DNA analyses reveal four genetically separated breeding units of the swordfish. Journal of Fish Biology 56: 1087-1098. Clark, J.M. (1988). Novel non-templated nucleotide addition reactions catalyzed by procaryotic and eucaryotic DNA polymerases. Nucleic Acids Research 16(20): 9677-9686. Clarke, K.R. and R.N. Gorley (2002). PRIMER v.5 User Manual/Tutorial, PRIMER-E, Plymouth. Clement, M., D. Posada and K.A. Crandall (2000). TCS: a computer program to estimate gene genealogies. Molecular Ecology 9: 1657-1659. CLIMAP (1976). The surface of ice-age Earth. Science 191(4232): 1131-1137. Coddington, J.A. and N. Scharff (1996). Problems with "soft" polytomies. Cladistics 12: 139-145. Coleman, M.A., M. Roughan, H.S. Macdonald, S.D. Connell, B.M. Gillanders, B.P. Kelaher and P.D. Steinberg (2011). Variation in the strength of continental boundary currents determines continent-wide connectivity in kelp. Journal of Ecology 99: 1026-1032. Colgan, D.J. and J.R. Paxton (1997). Biochemical genetics and recognition of a western stock of the common gemfish, Rexea solandri (Scombroidea, Gempylidae), in Australia. Marine and Freshwater Research 48: 103-118. Colson, I. and R.N. Hughes (2007). Contrasted patterns of genetic variation in the dogwhelk Nucella lapillus along two putative post-glacial expansion routes. Marine Ecology Progress Series 343: 183-191. Cooke, G.M. and L.B. Beheregaray (2007). Extremely high variability in the S72 intron of the Amazonian cardinal tetra (Paracheirodon axelrodi). Journal of Fish Biology 71(Supplement A): 132-140. Cooper, A., C. Lalueza-Fox, S. Anderson, A. Rambaut, J. Austin and R. Ward (2001). Complete mitochondrial genome sequences of two extinct moas clarify ratite evolution. Nature 409: 704-707. Cornuet, J.M. and G. Luikart (1996). Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001-2014. Corrigan, S. and L. Beheregaray (2009). A recent shark radiation: Molecular phylogeny, biogeography and speciation of wobbegong sharks (family: Orectolobidae). Molecular Phylogenetics and Evolution 52: 205-216. Craig, M.T., P.A. Hastings and D.J. Pondella (2004). Speciation in the Central American Seaway: the importance of taxon sampling in the identification of transisthmian geminate pairs. Journal of Biogeography 31: 1085-1091. Crandall, E.D., M.A. Frey, R.K. Grosberg and P.H. Barber (2008). Contrasting demographic history and phylogeographical patterns in two Indo-Pacific gastropods. Molecular Ecology 17: 611-626. Creer, S. (2007). Choosing and using introns in molecular phylogenetics. Evolutionary Bioinformatics 3: 99-108. Cunningham, C.W. (1997a). Is congruence between data partitions a reliable predictor of phylogenetic accuracy? Systematic Biology 46: 464-478. Cunningham, C.W. (1997b). Can three incongruence tests predict when data should be combined? Molecular Biology and Evolution 14: 733-740. Cuvier, G. (1829). In G. Cuvier, G. and A. Valenciennes. Histoire naturelle des poissons. Tome troisième. Suite du Livre troisième. Des percoïdes à dorsale unique 197
References à sept rayons branchiaux et à dents en velours ou en cardes. Histoire naturelle des poissons.: v. 3: i-xxviii + 2 pp. + 1-500, Pls. 41-71. Daemen, E., T. Cross, F. Ollevier and F.A.M. Volckaert (2001). Analysis of the genetic structure of European eel (Anguilla anguilla) using microsatellite DNA and mtDNA markers. Marine Biology 139: 755-764. Dannewitz, J., G.E. Maes, L. Johansson, H. Wickström, F.A.M. Volckaert and T. Järvi (2005). Panmixia in the European eel: a matter of time... Proceedings of the Royal Society B 272: 1129-1137. Dartnall, A.J. (1974). Littoral biogeography. Biogeography and ecology in Tasmania. W.D. Williams. The Hague, Junk: pp. 171-194. Davidson, J.F., R. Fox, D.D. Harris, S. Lyons-Abbott and L.A. Loeb (2003). Insertion of the T3 DNA polymerase thioredoxin binding domian enhances the processivity and fidelity of Taq DNA polymerase. Nucleic Acids Research 31(16): 4702-4709. Dawson, M.N. (2005). Incipient speciation of Catostylus mosaicus (Scyphozoa, Rhizostomeae, Catostylidae), comparative phylogeography and biogeography in south-east Australia. Journal of Biogeography 32: 515-533. de Queiroz, K. (1998). The general lineage concept of species, species criteria, and the process of speciation: A conceptual unification and terminological recommendations. Endless Forms. D.J. Howard and S.H. Berlocher. New York, Oxford University Press: pp. 57-75. de Queiroz, K. (2007). Species concepts and species delimitation. Systematic Biology 56(6): 879-886. DeLigny, W. and E.M. Pantelouris (1973). Origin of the European eel. Nature 246: 518519. Doiuchi, R. and T. Nakabo (2006). Molecular phylogeny of the stromateoid fishes (Teleostei: Perciformes) inferred from mitochondrial DNA sequences and compared with morphology-based hypotheses. Molecular Phylogenetics and Evolution 39: 111-123. Domingues, V.S., S. Stefanni, A. Brito, R.S. Santos and V.C. Almada (2008). Phylogeography and demography of the Blenniid Parablennius parvicornis and its sister species P. sanguinolentus from the northeastern Atlantic Ocean and the western Mediterranean Sea. Molecular Phylogenetics and Evolution 46: 397-402. DPI (2008). Fisheries Victoria Commercial Fish Production Information Bulletin 2008. Unpublished Report. Fisheries Victoria, Department of Primary Industries, Queenscliffe., Queenscliff, Victoria. pp. Drew, J., L.A. Rocha, P. Barber and M. Erdmann (2009). Assymetrical gene flow among populations of checkerboard wrasse from the Indian and Pacific Oceans. Combined 8th Indo Pacific Fish Conference and 2009 Australian Society of Fish Biologists Workshop and Conference. Fremantle, Western Australia. Drummond, A. and A. Rambaut (2007). BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 7: 214-221. Duffy, C.A.J. and C. Petherick (1999). A new size record for kahawai (Arripis trutta) from New Zealand. New Zealand Journal of Marine and Freshwater Research 33: 565-569. Duke, N.C., J.A.H. Benzie, J.A. Goodall and E.R. Ballment (1998). Genetic structure and evolution of species in the mangrove Avicennia (Avicenniaceae) in the IndoWest Pacific. Evolution 52: 1612-1626. Edgar, G.J. (2000). Australian Marine Life: The Plants and Animals of Temperate Waters. Frenchs Forest, Reed New Holland. 198
References Edwards, S.V. (2009). Is a new and general theory of molecular systematics emerging? Evolution 63(1): 1-19. Egglestone, D. (1975). Determination of age of kahawai Arripis trutta (Bloch & Schneider). New Zealand Journal of Marine and Freshwater Research 9(3): 293-8. Ellegren, H. (2004). Microsatellites: simple sequences with complex evolution. Nature Reviews Genetics 5: 435-445. Epicentre (2003). MasterPureTM DNA Purification Kit: Instructions for Use. Eschmeyer, W.N. (2011). Catalog of Fishes (electronic version). Estoup, A., P. Jarne and J.M. Cornuet (2002). Homoplasy and mutation model at microsatellite loci and their consequence for population genetics analysis. Molecular Ecology 11: 1591-1604. Excoffier, L., G. Laval and S. Schneider (2005). ARLEQUIN ver. 3.0: An integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1: 47-50. Excoffier, L. and H. Schneider (1999). Why hunter-gatherer populations do not show signs of Pleistocene demographic expansions. Proceedings of the National Academy of Sciences 96(19): 10597-10602. Eytan, R.I. and M.E. Hellberg (2010). Nuclear and mitochondrial sequence data reveal and conceal different demographic histories and population genetic processes in Caribbean reef fishes. Evolution 64: 3380-3397. Fairbridge, W.S. (1950). Some populations of the Australian "Salmon". Indo-Pacific Fisheries Council Proceedings, Cronulla, NSW. Fairclough, D.V., W.F. Dimmlich and I.C. Potter (2000a). Reproductive biology of the Australian herring Arripis georgiana. Marine and Freshwater Research 51: 619630. Fairclough, D.V., W.F. Dimmlich and I.C. Potter (2000b). Length and age compositions and growth rates of the Australian herring Arripis georgiana in different regions. Marine and Freshwater Research 51: 631-640. Farris, J.S., M. Källersjö, A.G. Kluge and C. Bult (1994). Testing the significance of incongruence. Cladistics 10: 315-319. Farris, J.S., M. Källersjö, A.G. Kluge and C. Bult (1995). Constructing a significance test for incongruence. Systematic Biology 44: 570-572. Fauvelot, C., G. Bernardi and S. Planes (2003). Reductions in the mitochondrial diversity of coral reef fish provide evidence of population bottlenecks resulting from Holocene sea-level change. Evolution 57: 1571-1583. Fauvelot, C., C. Lemaire, S. Planes and F. Bonhomme (2007). Inferring gene flow in coral reef fishes from different molecular markers: which loci to trust? Heredity 99: 331-339. Felsenstein, J. (1973). Maximum likelihood and minimum-steps methods for estimating evolutionary trees from data in discrete characters. Systematic Zoology 22(3): 240249. Felsenstein, J. (1981). Evolutionary trees from DNA sequences: A maximum likelihood approach. Journal of Molecular Evolution 17(6): 368-376. Felsenstein, J. (1985). Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783-791. Felsenstein, J. (1988). Phylogenies from molecular sequences: Inference and reliability. Annual Review of Genetics 22: 521-565. Ferguson, J.W.H. (2002). On the use of genetic divergence for identifying species. Biological Journal of the Linnean Society 75: 509-516. 199
References Fernando, P., B.J. Evans, J.C. Morales and D.J. Melnick (2001). Electrophoresis artifacts - a previously unrecognized cause of error in microsatellite analysis. Molecular Ecology Notes 1: 325-328. Fischer, M.C., M. Foll, L. Excoffier and G. Heckel (2011). Enhanced AFLP genome scans detect local adaptation in high-altitude populations of a small rodent (Microtus arvalis). Molecular Ecology 20(7): 1450-1462. Fitch, W.M. (1977). On the problem of discovering the most parsimonious tree. American Naturalist 111: 223-257. Forster (1801). In M.E. Bloch and J.G. Schneider. Systema Ichthyologiae iconibus cx illustratum. Post obitum auctoris opus inchoatum absolvit, correxit, interpolavit Jo. Gottlob Schneider, Saxo. Berolini. M.E. Blochii, Systema Ichthyologiae, Sumtibus Auctoris Impressum et Bibliopolio Sanderiano Commissum. i-lx + 1-584 pp. Foster, E.G., D.A. Ritz, J.E. Osborn and K.M. Swadling (2001). Schooling affects the feeding success of Australian salmon (Arripis trutta) when preying on mysid swarms (Paramesopodopsis rufa). Journal of Experimental Marine Biology and Ecology 261: 93-106. Fowler, A.J., R. McGarvey, M. Steer and J.E. Feensta (2008). The South Australian marine scalefish fishery stock status report. Unpublished Report. SARDI publication number F2007/ 000565-3, Adelaide. pp. Fraser, C.I., H.G. Spencer and J.M. Waters (2009). Glacial oceanographic contrasts explain phylogeography of Australian bull kelp. Molecular Ecology 18: 2287-2296. Friesen, V.L. (2000). Introns. Molecular Methods in Ecology. A.J. Baker. Oxford, Blackwell Science. Fu, Y.X. (1997). Statistical tests of neutrality against population growth, hitchhiking and background selection. Genetics 147: 915-925. Fumigalli, L., P. Taberlet, L. Favre and J. Hausser (1996). Origin and evolution of homologous repeated sequences in the mitochondrial DNA control region of shrew. Molecular Biology and Evolution 13: 31-46. Gadagkar, S.R., M.S. Rosenberg and S. Kumar (2005). Inferring species phylogenies from multiple genes: concatenated sequence tree versus consensus gene tree. Journal of Experimental Zoology 304B: 64–74. Gales, R., D. Pemberton, M. Clarke and C.C. Lu (1992). Stomach contents of longfinned whales (Globicephala melas) and bottlenose dolphins (Tursiops truncatus) in Tasmania. Marine Mammal Science 8: 405-413. Galtier, N., B. Nabholz, S. Glémin and G.D.D. Hurst (2009). Mitochondrial DNA as a marker of molecular diversity: a reappraisal. Molecular Ecology 18: 4541–4550. Gavrilets, S. (2003). Models of speciation: what have we learned in 40 years? Evolution 57: 2197-2215. Gehring, R.E. (1978). Basic Behavioural Statistics. Boston, Houghton Mifflin Company. Gemmell, N.J., P.S. Western, J.M. Watson and J.A.M. Graves (1996). Evolution of mammalian mitochondrial control region – comparisons of control region sequences between Monotreme and Therian mammals. Molecular Biology and Evolution 13: 798-808. Gillanders, B.M., K.W. Able, J.A. Brown, D.B. Egglestone and P.F. Sheridan (2003). Evidence of connectivity between juvenile and adult habitats for mobile marine fauna: an important component of nurseries. Marine Ecology Progress Series 247: 281-295. 200
References Gingele, F.X., P. De Deckker and C.-D. Hillenbrand (2001). Late Quaternary fluctuations of the Leeuwin Current and palaeoclimates on the adjacent land masses: clay mineral evidence. Australian Journal of Earth Sciences 48(6): 867874. Giribet, G. and W.C. Wheeler (1999). On gaps. Molecular Phylogenetics and Evolution 13: 132-143. Godfrey, J.S. and K.R. Ridgway (1985). The large-scale environment of the polewardflowing Leeuwin Current, Western Australia: longshore steric height gradients, wind stresses and geostrophic flow. Journal of Physical Oceanography 15: 481– 495. Gold, J.R., E. Pak and D.A. DeVries (2002). Population structure of king mackerel (Scomberomorus cavalla) around peninsular Florida, as revealed by microsatellite DNA. Fishery Bulletin 100: 491-509. Goldstein, P.Z., R. DeSalle, G. Amato and A.P. Vogler (2000). Conservation genetics at the species boundary. Conservation Biology 14(1): 120-131. Gomon, M.F., D. Bray and R.H. Kuiter (2008). Fishes of Australia's Southern Coast. Frenchs Forest, New Holland Publishers. Gomon, M.F., J.C.M. Glover and R.H. Kuiter (1994). The Fishes of Australia's South Coast. Adelaide, State Print. Gomulski, L.M., K. Bourtzis, S. Brogna, P.A. Morandi, C. Bonvicini, F. Sebastiani, C. Torti, C.R. Guglielmino, C. Savakis, G. Gasperi and A.R. Malacrida (1998). Intron size polymorphism of the Adh1 gene parallels the worldwide colonization history of the Mediterranean fruit fly, Ceratitis capitata. Molecular Ecology 7: 1729-1741. Goudet, J. (1995). FSTAT (Version 1.2): A computer program to calculate F-statistics. Heredity 86(6): 485-486. Grant, W.S. and B.W. Bowen (1998). Shallow population histories in deep evolutionary lineages of marine fishes: Insights from sardines and anchovies and lessons for conservation. Journal of Heredity 89: 415-426. Gromek, K. and T. Kaczorowski (2005). DNA sequencing by indexer walking. Clinical Chemistry 51(9): 1612-1618. Guinand, B. (1996). Use of a multivariate model using allele frequency distributions to analyse patterns of genetic differentiation among populations. Biological Journal of the Linnean Society 58: 173-195. Guindon, S. and O. Gascuel (2003). A simple, fast and accurate method to estimate large phylogenies by maximum-likelihood. Systematic Biology 52: 696-704. Guo, S.W. and E.A. Thompson (1992). Performing the Exact test of Hardy-Weinberg proportion for multiple alleles. Biometrics 48: 361-372. Hale, M.L., P.W.W. Lurz, M.D.F. Shirley, S. Rushton, R.M. Fuller and K. Wolff (2001). Impact of landscape management on the genetic structure of red squirrel populations. Science 293: 2246-2248. Haney, R.A., B.R. Silliman and D.M. Rand (2007). A multi-locus assessment of connectivity and historical demography in the bluehead wrasse (Thalassoma bifasciatum). Heredity 98: 294-302. Hare, M.P. (2001). Prospects for nuclear gene phylogeography. Trends in Ecology and Evolution 16(12): 700-706. Hare, M.P., F. Cipriano and S.R. Palumbi (2002). Genetic evidence on the demography of speciation in allopatric dolphin species. Evolution 56(4): 804-816. Harpending, H. (1994). Signature of ancient population growth in a low-resolution mitochondrial mismatch distribution. Human Biology 66: 591-600. 201
References Hassan, M., C. Lemaire, C. Fauvelot and F. Bonhomme (2002). Seventeen new exonprimed intron-crossing polymerase chain reaction amplifiable introns in fish. Molecular Ecology Notes 2: 334-340. Hebert, P.D., M.Y. Stoeckle, T.S. Zemlak and C.M. Francis (2004). Identification of birds through DNA barcodes. PLoS Biology 2(10): 1657-1663. Hedgecock, D. (1994). Does variance in reproductive success limit effective population sizes of marine organisms? Genetics and evolution of aquatic organisms. A.R. Beaumont, Chapman and Hall, London: 122–134. Hedrick, P.W. (2005). A standardized genetic differentiation measure. Evolution 59: 1633-1638. Heled, J. and A. Drummond (2010). Bayesian inference of species trees from multilocus data. Molecular Biology and Evolution 27(3): 570–580. Hennig, W. (1965). Phylogenetic systematics. Annual Review of Entomology 10: 97116. Henry, G.W. and J.M.E. Lyle (2003). The National Recreational and Indigenous Fishing Survey. Unpublished Report. Australian Government Department of Agriculture, Fisheries and Forestry. pp. Hewitt, G. (2000). The genetic legacy of the Quaternary ice ages. Nature 405: 907-913. Hey, J. (1991). The structure of genealogies and the distribution of fixed differences between DNA sequence samples from natural populations. Evolution 128: 831-840. Hey, J. (2006a). Recent advances in assessing gene flow between diverging populations and species. Current Opinions in Genetics and Development 16: 592-596. Hey, J. (2006b). On the failure of modern species concepts. Trends in Ecology and Evolution 21(8): 447-450. Hey, J. and J. Wakeley (1997). A coalescent estimator of the population recombination rate. Genetics 145: 833-846. Hickerson, M.J. and C.W. Cunningham (2005). Contrasting quaternary histories in an ecologically divergent sister pair of low-dispersing intertidal fish (Xiphister) revealed by multilocus DNA analyses. Evolution 59: 344-360. Hidas, E.Z., T.L. Costa, D.J. Ayre and T.E. Minchinton (2007). Is the species composition of rocky intertidal invertebrates across a biogeographic barrier in south-eastern Australia related to their potential for dispersal? . Marine and Freshwater Research 58: 835-842. Hillis, D.M., J.P. Huelsenbeck and C.W. Cunningham (1994). Application and accuracy of molecular phylogenies. Science 264: 671-677. Hoedt, F.E. and W.F. Dimmlich (1994). Diet of subadult Australian Salmon, Arripis truttaceus, in Western Port, Victoria. Australian Journal of Marine and Freshwater Research 45: 617-623. Hoelzer, G.A. and D.J. Melnick (1994). Patterns of speciation and limits to phylogenetic resolution. Trends in Ecology and Evolution 9: 104-107. Hoffman, J.I. and W. Amos (2005). Microsatellite genotyping errors: detection approaches, common sources and consequences for paternal exclusion. Molecular Ecology 14: 599–612. Hoolihan, J.P., P. Anandh and L. van Herwerden (2006). Mitochondrial DNA analyses of narrow-barred Spanish mackerel (Scomberomorus commerson) suggest a single genetic stock in the ROPME sea area (Arabian Gulf, Gulf of Oman, and Arabian Sea). ICES Journal of Marine Science 63: 1066-1074.
202
References Hostetler, S., N. Pisias and A. Mix (2006). Sensitivity of Last Glacial Maximum climate to uncertainties in tropical and subtropical ocean temperatures. Quaternary Science Reviews 25(11-12): 1168-1185. Howard, W.R. and W.L. Prell (1992). Late Quaternary surface circulation of the southern Indian Ocean and its relationship to orbital variations. Paleoceanography 7(1): 79-117. Hubert, N., F. Duponchelle, J. Nuñez, R. Rivera and J.-F. Renno (2006). Evidence of reproductive isolation among closely related species of Serrasalmus (Ostariophysii, Characidae) from the Upper Madeira River, Amazon, Bolivia. Journal of Fish Biology 69(Supplement A): 31-51. Huelsenbeck, J.P., J.J. Bull and C.W. Cunningham (1996). Combining data in phylogenetic analysis. Trends in Ecology and Evolution 11: 152-158. Huelsenbeck, J.P. and K.A. Crandall (1997). Phylogeny estimation and hypothesis testing using maximum likelihood. Annual Review of Ecology and Systematics 28: 437-466. Huelsenbeck, J.P., B. Larget, R.E. Miller and F. Ronquist (2002). Potential applications and pitfalls of Bayesian inference of phylogeny. Systematic Biology 51(5): 673688. Hughes, J.M., J. Stewart, B.M. Gillanders and I.M. Suthers (2011). An examination of the population structure of the eastern Australian salmon (Arripis trutta) stock in south-eastern Australia. Proceedings of the International Symposium on Advances in Fish Tagging and Marking Technology. February 2008, Auckland, New Zealand. Transactions of the American Fisheries Society 76. Hull, J.M., D.P. Mindell, S.L. Talbot, E.H. Kay, H.E. Hoeskstra and H.B. Ernest (2010). Population structure and plumage polymorphism: The intraspecific evolutionary relationships of a polymorphic raptor, Buteo jamaicensis harlani. BMC Evolutionary Biology 10: 224. Hurt, C., A. Anker and N. Knowlton (2009). A multilocus test of simultaneous divergence across the Isthmus of Panama using snapping shrimp in the genus Alpheus. Evolution 63(2): 514-530. Huson, D.H. (1998). SplitsTree: a program for analyzing and visualizing evolutionary data. Bioinformatics 14: 68-73. Huson, D.H. and D. Bryant (2006). Application of phylogenetic networks in evolutionary studies. Molecular Biology and Evolution 23: 254-267. Hutchins, J.B. (1987). Description of a new plesiopid fish from south-western Australia, with a discussion of the zoogeography of Paraplesiops. Records of the Western Australian Museum 13: 231-240. Hutchins, J.B. (1994). A survey of the nearshore reef fish fauna of Western Australia's west and south coasts - The Leeuwin Province. Records of the Western Australian Museum Supplement 46: 1-66. Hutchins, J.B. and R. Swainston (1999). Sea Fishes of Southern Australia. Perth, Swainston Publishing. Hutson, K.S., C.A. Styan, I. Beveridge, M.J. Keough, X. Zhu, Y. Abs EL-Osta and R.B. Gasser (2004). Elucidating the ecology of bucephalid parasites using a mutation scanning approach. Molecular and Cellular Probes 18: 139-146. Inoue, J.G., M. Miya, B. Venkatesh and M. Nishida (2005). The mitochondrial genome of Indonesian coelocanth Latimeria menadoensis (Sarcopterygii: Coelocanthiformes) and divergence time estimation between the two coelocanths. Gene 349: 227-235. 203
References Janko, K., G. Lecointre, A. DeVries, A. Couloux, C. Cruaud and C. Marshall (2007). Did glacial advances during the Pleistocene influence differently the demographic histories of benthic and pelagic Antarctic shelf fishes? – Inferences from intraspecific mitochondrial and nuclear DNA sequence diversity. BMC Evolutionary Biology 7: 220-236. Jansen, E., J. Overpeck, K.R. Briffa, J.-C. Duplessy, F. Joos, V. Masson-Delmotte, D. Olago, B. Otto-Bliesner, W.R. Peltier, S. Rahmstorf, R. Ramesh, D. Raynaud, D. Rind, O. Solomina, R. Villalba and D. Zhang (2007). Palaeoclimate. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. S. Solomon, D. Qin, M. Manninget al. Cambridge, United Kingdom and New York, NY, USA., Cambridge University Press. Jarne, P. and P.J.L. Lagoda (1996). Microsatellites, from molecules to populations and back. Trends in Ecology and Evolution 11: 424-429. Johns, G.C. and J.C. Avise (1998). A comparative summary of genetic distances in the vertebrates from the mitochondrial cytochrome b gene. Molecular Biology and Evolution 15(11): 1481-1490. Johnson, M.S. and R. Black (1984). Pattern beneath the chaos: the effect of recruitment on genetic patchiness in an intertidal limpet. Evolution 38(6): 1371-1383. Johnson, N.K. and C. Cicero (2004). New mitochondrial DNA data affirm the importance of Pleistocene speciation in North American birds. Evolution 58: 11221130. Jones, A.G., G.I. Moore, C. Kvarnemo, D. Walker and J.C. Avise (2003). Sympatric speciation as a consequence of male pregnancy in seahorses. Proceedings of the National Academy of Sciences 100(11): 6598-6603. Jordan, D.S. (1908). The law of the geminate species. American Naturalist 42: 73-80. Kassahn, K.S., S.C. Donnellan, A.J. Fowler, K.C. Hall, M. Adams and P.W. Shaw (2003). Molecular and morphological analyses of the cuttlefish Sepia apama indicate a complex population structure. Marine Biology 143: 947-962. Kaustuv, R., D. Jablonski and J.W. Valentine (2001). Climate change, species range limits and body size in marine bivalves. Ecology Letters 4: 366-370. Kendall, M. (1970). Rank Correlation Methods. London, Griffin. Kennedy, M., B.R. Holland, R.D. Gray and H.G. Spencer (2005). Untangling long branches: Identifying conflicting phylogenetic signals using spectral analysis, NeighborNet, and consensus networks. Systematic Biology 54(4): 620-633. Ketmaier, V. and C. Bernardini (2005). Structure of the mitochondrial control region of the Eurasian Otter (Lutra lutra; Carnivora, Mustelidae): patterns of genetic heterogeneity and implications for conservation of the species in Italy. Journal of Heredity 96(4): 318-328. Kilpatrick, C.W. (2002). Noncryogenic preservation of mammalian tissues for DNA extraction: an assessment of storage methods. Biochemical Genetics 40(1/2): 53-62. Kimura, M. and J.F. Crow (1964). The number of alleles that can be maintained in a finite population. Genetics 49: 725-738. Kirkwood, G.P. and M.H. Walker (1984). A new method for estimating tag shedding rates, with application to data for Australian Salmon, Arripis trutta esper Whitley. Australian Journal of Marine and Freshwater Research 35: 601-606. Kishino, H., T. Miyata and M. Hasegawa (1990). Maximum likelihood inference of protein phylogeny and the origin of chloroplasts. Journal of Molecular Evolution 31: 151-160. 204
References Kliman, R.M., P. Andolfatto, J.A. Coyne, F. Depaulis, M. Kreitman, A.J. Berry, J. McCarter, J. Wakeley and J. Hey (2000). The population genetics of the origin and divergence of the Drosophila simulans complex species. Genetics 156: 1913-1931. Knowles, L.L. (2008). Why does a method that fails continue to be used? Evolution 62: 2713-2717. Knowles, L.L. (2009). Statistical Phylogeography. Annual Review of Ecology and Systematics 40: 593-612. Knowles, L.L. and B.C. Carstens (2007). Delimiting species without monophyletic gene trees. Systematic Biology 56: 887-895. Knowlton, N. and L.A. Weigt (1998). New dates and new rates for divergence across the Isthmus of Panama. Proceedings of the Royal Society B 265: 2257-2263. Knutsen, H., P.E. Jorde, C. André and N.C. Stenseth (2003). Fine-scaled geographical population structuring in a highly mobile marine species: the Atlantic cod. Molecular Ecology 12: 385-394. Kocher, T.D., K. Thomas, A. Meyer, S.V. Edwards, S. Paabo, F.X. Villablanca and A.C.C. Wilson (1989). Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with conserved primers. Proceedings of the National Academy of Sciences 86: 6196-6200. Kong, X., J. Yu and L. Zhou (2007). Comparative analysis of 5'-end sequence of the mitochondrial control region of six flatfish species (Pleuronectidae) from the Yellow Sea. The Raffles Bulletin of Zoology Supplement 14: 111-120. Kraus, J., P.H. Dear, J.L. Pollack, M. Slatkin, H. Spriggs, I. Barnes, A.M. Lister, I. Ebersberger, S. Paabo and M. Hofreiter (2006). Multiplex amplification of the mammoth mitochondrial genome and the evolution of Elephantidae. Nature 439: 724-727. Lambeck, K., T.M. Esat and E.K. Potter (2002). Links between climate and sea levels for the past three million years. Nature 419: 199-206. Lavoué, S., J.P. Sullivan and C.D. Hopkins (2003). Phylogenetic utility of the first two introns of the S7 ribosomal protein gene in African electric fishes (Mormyroidea: Teleostei) and congruence with other molecular markers. Biological Journal of the Linnean Society 78: 273-292. Lawrence, K.T. and T.D. Herbert (2005). Late Quaternary sea-surface temperatures in the western Coral Sea: Implications for the growth of the Australian Great Barrier Reef. Geology 33(8): 677-680. Leach, B.F., J.M. Davidson, L.M. Horwood and S. Mallon (1996). The estimation of live fish size from archaeological cranial bones of the New Zealand kahawai Arripis trutta. Tuhinga: Records of the Museum of New Zealand Te Papa Tongarewa 7: 1-20. Leaché, A.D., M.S. Koo, C.L. Spencer, T.J. Papenfuss, R.N. Fisher and J.A. McGuire (2009). Quantifying ecological, morphological, and genetic variation to delimit species in the coast horned lizard species complex (Phrynosoma). Proceedings of the National Academy of Sciences 106(30): 12418-12423. Leberg, P.L. (2002). Estimating allelic richness: effects of sample size and bottlenecks. Molecular Ecology 11: 2445-2449. Lee, W.J., J. Conroy, W. Huntting Howell and T.D. Kocher (1995). Structure and evolution of teleost mitochondrial control regions. Journal of Molecular Evolution 41: 54-66.
205
References Legendre, P. (2005). Species associations: the Kendall coefficient of concordance revisited. Journal of Agricultural, Biological, and Environmental Statistics 10(2): 226-245. Lenanton, R. (1978). Age, spawning time, and fecundity of Australian Herring (Arripis georgianus C. & V.) (Pisces: Arripidae) from the waters around Rottnest Island, Western Australia. Australian Journal of Marine and Freshwater Research 29: 599-612. Lenanton, R. (1982). Alternative non-estuarine nursery habitats for some commercially and recreationally important fish species of south-western Australia. Australian Journal of Marine and Freshwater Research 33: 881-900. Lenanton, R., L. Joll, J. Penn and K. Jones (1991). The influence of the Leeuwin Current on coastal fisheries of Western Australia. Journal of the Royal Society of Western Australia 74: 101-114. Lenanton, R.C.J. and N.G. Hall (1976). The Western Australian amateur fishery for Australian Herring (Arripis georgianus): results of the 1973 creel census. Unpublished Report. Department of Fisheries and Wildlife, Western Australia. 59 pp. pp. Lessa, E.P. (1992). Rapid surveying of DNA sequence variation in natural populations. Molecular Biology and Evolution 9(2): 323-330. Lessios, H.A. (1998). The first stage of speciation as seen in organisms separated by the Isthmus of Panama. Endless forms: Species and speciation. D. Howard and S. Berlocher. Oxford, Oxford University Press: pp. 186-201. Lessios, H.A. (2008). The Great American Schism: Divergence of marine organisms after the rise of the Central American Isthmus. Annual Review of Ecology, Evolution and Systematics 39: 63-91. Lessios, H.A., M.J. Garrido and B.D. Kessing (2001). Demographic history of Diadema antillarum, a keystone herbivore on Caribbean reefs. Proceedings of the Royal Society B 268: 583-588. Lewis, P.O. (2001). A likelihood approach to estimating phylogeny from discrete morphological character data. Systematic Biology 50: 913-925. Li, W. and M. Gouy (1991). Statistical methods for testing molecular phylogenies. Phylogenetic Analysis of DNA Sequences. M.M. Miyamoto and J. Cacraft. Oxford, Oxford University Press: pp.249-277. Li, Y.-C., A.B. Korol, T. Fahima, A. Beiles and E. Nevo (2002). Microsatellites: genomic distribution, putative functions and mutational mechanisms: a review. Molecular Ecology 11: 2453-2465. Lintas, C., J. Hirano and S. Archer (1998). Genetic variation of the European eel (Anguilla anguilla). Molecular Marine Biology and Biotechnology 7: 263-269. Lohman, D.J., D. Peggie, N.E. Pierce and R. Meier (2008). Phylogeography and genetic diversity of a widespread Old World butterfly, Lampides boeticus (Lepidoptera: Lycaenidae). BMC Evolutionary Biology 8: 301-314. López, M.D., M.U. Alcocer and P.D. Jaimes (2010). Phylogeography and historical demography of the Pacific Sierra mackerel (Scomberomorus sierra) in the Eastern Pacific. BMC Genetics 11: 34-45. Luikart, G. and J.M. Cornuet (1998). Empirical evaluation of a test for identifying recently bottlenecked populations from allele frequency data. Conservation Biology 12(1): 228-237.
206
References Lunt, D.H., L.E. Whipple and B.C. Hyman (1998). Mitochondrial DNA variable number tandem repeats (VNTRs): utility and problems in molecular ecology. Molecular Ecology 7: 1441-1455. Lynch, M. (2002). Intron evolution as a population-genetic process. Proceedings of the National Academy of Sciences 99(9): 6118-6123. MacDonald, C.M. (1980). Population Structure, Biochemical Adaptation and Systematics in Temperate Marine Fishes of the Genera Arripis and Chrysophrys (Pisces: Perciformes). Canberra, Australian National University: 336. MacDonald, C.M. (1983). Taxonomic and evolutionary studies on marine fishes of the genus Arripis (Perciformes: Arripidae). Bulletin of Marine Science 33(3): 780. Maddison, W.P. (1997). Gene trees in species trees. Systematic Biology 46(523-536). Maddison, W.P., M.J. Donoghue and D.R. Maddison (1984). Outgroup analysis and parsimony. Systematic Zoology 33(1): 83-103. Maes, G.E. and F.A.M. Volckaert (2002). Clinal genetic variation and isolation by distance in the European eel Anguilla anguilla (L.). Biological Journal of the Linnean Society 77: 509-521. Malcolm, W.B. (1959). The populations of Australian "Salmon", Arripis trutta (Bloch & Schneider), in Australian waters. Australian Journal of Marine and Freshwater Research 10: 22-29. Malcolm, W.B. (1960). Area of distribution, and movement of the western subspecies of the Australian "Salmon", Arripis trutta esper Whitley. Australian Journal of Marine and Freshwater Research 11: 282-325. Mallarino, R., E. Bermingham, K.R. Willmott, A. Whinnett and C.D. Jiggins (2005). Molecular systematics of the butterfly genus Ithomia (Lepidoptera: Ithominae): a composite phylogenetic hypothesis based on seven genes. Molecular Phylogenetics and Evolution 34: 625-644. Mallet, J. (2005). Hybridization as an invasion of the genome. Trends in Ecology and Evolution 20: 229-237. Mallet, J. (2007). Species, concepts of. Encyclopedia of Biodiversity. S.A. Levin. Oxford, Elsevier: pp. 1-15. Mallet, J. and K. Willmott (2003). Taxonomy: Renaissance or Tower of Babel? Trends in Ecology and Evolution 18: 57-59. Marchant, S. and P.J. Higgins, (Eds.) (1990). Handbook of Australian, New Zealand and Antarctic Birds. Volume 1: Ratites to Ducks.: 1400. Marko, P.B. (2002). Fossil calibration of molecular clocks and the divergence times of geminate species pairs separated by the Isthmus of Panama. Molecular Biology and Evolution 19: 2005-2021. Marko, P.B. and A.L. Moran (2009). Out of sight, out of mind: high cryptic diversity obscures the identities and histories of geminate species in the marine bivalve subgenus Acar. Journal of Biogeography 36: 1861-1880. Marko, P.B., L. Rogers-Bennett and A.B. Dennis (2007). MtDNA population structure and gene flow in lingcod (Ophiodon elongatus): limited connectivity despite longlived pelagic larvae. Marine Biology 150: 1301-1311. Maruyama, T. and P.A. Fuerst (1985). Population bottlenecks and non-equilibrium models in population genetics. II. Number of alleles in a small population that was formed by a recent bottleneck. Genetics 111: 675-689. Mateus, C.S., P.R. Almeida, B.R. Quintella and M.J. Alves (2011). MtDNA markers reveal the existence of allopatric evolutionary lineages in the threatened lampreys 207
References Lampetra fluviatilis (L.) and Lampetra planeri (Bloch) in the Iberian glacial refugium Conservation Genetics: doi 10.1007/s10592-011-0209-x. Mayden, R.L. (1999). Consilience and a hierarchy of species concepts: advances toward closure on the species puzzle. Journal of Nematology 31: 95-116. McCartney, M.A. and H.A. Lessios (2002). Quantitative analysis of gametic incompatibility between closely related species of neotropical sea urchins. Biological Bulletin 202: 166-181. McCormack, S. (2000). The fishery for Australian salmon: management issues. Fish Movement and Migration. Australian Society for Fish Biology Workshop Proceedings, Bendigo, Victoria, 28-29 September 1999, Australian Society for Fish Biology, Sydney. McGaughran, A., G. Torricelli, A. Carapelli, F. Frati, M.I. Stevens, P. Convey and I.D. Hogg (2010). Contrasting phylogeographical patterns for springtails reflect different evolutionary histories between the Antarctic Peninsula and continental Antarctica. Journal of Biogeography 37(1): 103-119. Meirmans, P.G. (2007). Using the AMOVA framework to estimate a standardized genetic differentiation measure. Evolution 60: 2399-2402. Metzker, M.L. (2010). Sequencing technologies - the next generation. Nature Reviews Genetics 11: 31-46. Meyer, A., T.D. Kocher, P. Basasibwaki and A.C. Wilson (1990). Monophyletic origin of Lake Victoria cichlid fishes suggested by mitochondrial DNA sequences. Nature 347: 550-553. Meyer, C. and G. Paulay (2005). DNA barcoding: error rates based on comprehensive sampling. PLoS Biology 3: e422. Miller, B. (2007). Native fish threatens commercial catch: interview with Malcolm Poole, Chairman, Recreational Fishing Alliance of NSW. Radio program 'AM'. Australian Broadcasting Corporation. Miura, O., M.E. Torchin and E. Bermingham (2010). Molecular phylogenetics reveals differential divergence of coastal snails separated by the Isthmus of Panama. Molecular Phylogenetics and Evolution 56: 40-48. Montague, T.L. and J.M. Cullen (1987). The diet of the little penguin Eudyptula minor at Phillip Island, Victoria. Emu 86(4): 138-149. Moritz, C. and C. Cicero (2004). DNA barcoding: promises and pitfalls. PLoS One 2(10): e354. Morrison, C.L., D.P. Lemarié, R.M. Wood and T.L. King (2006). Phylogeographic analyses suggest multiple lineages of Crystallaria asprella (Percidae: Etheostominae). Conservation Genetics 7: 129-147. Morrison, D.A. (2005). Networks in phylogenetic analysis: new tools for population biology. International Journal for Parasitology 35: 567-582. Moulton, V. and K.T. Huber (2009). Split networks. A tool for exploring complex evolutionary relationships in molecular data. The Phylogenetic Handbook. A practical approach to phylogenetic analysis and hypothesis testing. P. Lemey, M. Salemi and A.-M. Vandamme. Cambridge, Cambridge University Press: pp. 631653. Moyer, G.R., K.O. Winemiller, M.V. McPhee and T.F. Turner (2005). Historical demography, selection, and coalescence of mitochondrial and nuclear genes in Prochilodus species of northern South America. Evolution 59(3): 599-610.
208
References Muss, A., D.R. Robertson, C.A. Stepian, P. Wirz and B.W. Bowen (2001). Phylogeography of Ophioblennius: The role of ocean currents and geography in reef fish evolution. Evolution 55: 561-572. Nei, M. (1987). Molecular Evolutionary Genetics. New York, Columbia University Press. Nei, M. (1991). Relative efficiencies of different tree-making methods for molecular data. Phylogenetic Analysis of DNA Sequences. M.M. Miyamoto and J. Cacraft. Oxford, Oxford University Press: pp. 90-128. Nei, M. and S. Kumar (2000). Molecular Evolution and Phylogenetics. New York, Oxford University Press. Nei, M., T. Maruyama and R. Chakraborty (1975). The bottleneck effect and genetic variability in populations. Evolution 29: 1-10. Nei, M., F. Tajima and Y. Tateno (1983). Accuracy of estimated phylogenetic trees from molecular data. Journal of Molecular Evolution 19: 153-170. Nei, M. and N. Takezaki (1994). Estimation of genetic distances and phylogenetic trees from DNA analysis. Proceedings of the 5th World Congress of Genetics Applied to Livestock Production. Neira, F.J., A.G. Miskiewicz and B.D. Bruce (1997). Larvae of five fish families with pattern 10 of the ramus lateralis accessorius nerve (Arripididae, Girellidae, Kyphosidae, Microcanthidae and Scorpididae): relevance to relationships. Bulletin of Marine Science 60: 117-138. Nesbø, C.L., M.O. Arab and K.S. Jakobsen (1998). Heteroplasmy, length and sequence variation in the mtDNA control regions of three percid fish species (Perca fluviatilis, Acerina cernua, Stizostedion lucioperca). Genetics 148: 1907-1919. Nicholls, A.G. (1969). The Australian "Salmon" (Arripis trutta esper Whitley). The western subspecies fishery from 1961-1969. Unpublished Report. CSIRO Division of Fisheries and Oceanography, Cronulla, NSW. 64 pp. pp. Nicholls, A.G. (1971). A monitoring programme for Australian "Salmon" in Western Australia. Unpublished Report. CSIRO Division of Fisheries and Oceanography, Cronulla, NSW. 40 pp. pp. Nicholls, A.G. (1973). Growth in the Australian "Salmon" Arripis trutta (Bloch & Schneider). Australian Journal of Marine and Freshwater Research 24: 159-176. Nurthen, R.K., R.A. Cameron and D.A. Briscoe (1992). Population genetics of tailor, Pomatomus saltatrix (Linnaeus) (Pisces: Pomatomidae), in Australia. Australian Journal of Marine and Freshwater Research 43: 1481-1486. O’Hara, T.D. and G.C.B. Poore (2000). Patterns of distribution for southern Australian marine echinoderms and decapods. Journal of Biogeography 27: 1321-1335. Ohta, T. and M. Kimura (1973). A model of mutation appropriate to estimate the number of electrophoretically detectable alleles in a finite population. Genetics Research 22: 201-204. Ota, T. (1993). DISPAN (Genetic Distance and Phylogenetic Analysis), Pennsylvania State University. available at http://iubio.bio.indiana.edu/soft/molbio/. Page, R.D.M. and E.C. Holmes (1998). Molecular Evolution. A Phylogenetic Approach. Oxford, Blackwell Science. Paine, M.A., J.R. McDowell and J.E. Graves (2007). Specific identification of western Atlantic Ocean scombrids using mitochondrial DNA cytochrome c oxidase subunit I (COI) gene region sequences. Bulletin of Marine Science 80: 353-367.
209
References Palm, S., J. Dannewitz, T. Prestegaard and H. Wickström (2009). Panmixia in European eel revisited: no genetic difference between maturing adults from southern and northern Europe. Heredity 103: 82-89. Palumbi, S.R. and C.S. Baker (1994). Contrasting population structure from nuclear intron sequences and mtDNA of Humpback Whales. Molecular Biology and Evolution 11(3): 426-435. Parker, P.G., A.A. Snow, M.D. Schug, G.C. Booton and P.A. Fuerst (1998). What molecules can tell us about populations: choosing and using a molecular marker. Ecology 79(2): 361-382. Paulin, C. (1993). Review of the Australian fish family Arripididae (Percomorpha), with the description of a new species. Australian Journal of Marine and Freshwater Research 44: 459-471. Paulo, O.C., C. Dias, M.W. Bruford, W.C. Jordan and R.A. Nichols (2001). The persistence of Pliocene populations through the Pleistocene climatic cycles: evidence from the phylogeography of an Iberian lizard. Proceedings of the Royal Society B 268: 1625-1630. Percides, A. and M.M. Coelho (2006). Comparative phylogeography of Zacco platypus and Opsariichthys bidens (Teleostei: Cyprinidae) in China based on cytochrome b sequences. Journal of Zoological Systematics and Evolutionary Research 44(4): 330-338. Pereira, R., V. Brondani and D. Grattapaglia (2001). Cost-effective method to synthesize a fluorescent internal DNA standard for automated fragment sizing. BioTechniques 31: 793-800. Petit, J.R. (2008). The coup de grâce for the nested clade phylogeographic analysis? Molecular Ecology 17: 516-518. Petit, J.R., J. Jouzel, D. Raynaud, N.I. Barkov, J.-M. Barnola, I. Basile, M. Benders, J. Chappellaz, M. Davis, G. Delayque, M. Delmotte, V.M. Kotlyakov, M. Legrand, V.Y. Lipenkov, C. Lorius, L. Pépin, C. Ritz, E. Saltzman and M. Stievenard (1999). Climate and atmospheric history of the past 420,000 years from the Vostok ice core, Antarctica. Nature 399: 429-436. Petrusevics, P. and J.A.T. Bye (1995). The Leeuwin Current, wind, mean sea level and variability in Australian salmon larvae recruitment. Proceedings of the Ocean and Atmosphere Pacific International Conference, Adelaide, South Australia. Poe, S. and A.L. Chubb (2004). Birds in a bush: five genes indicate explosive evolution of avian orders. Evolution 58: 404-415. Pompanon, F., A. Bonin, E. Bellemain and P. Taberlet (2005). Genotyping errors: causes, consequences and solutions. Nature Reviews Genetics 6: 847-859. Ponder, W.F. (1995). Systematics, evolution & biogeography of species in coastal habitats. Coastal Marine Ecology of Temperate Australia. A.J. Underwood and M.G. Chapman. Sydney, University of New South Wales: pp. 16-27. Posada, D. (2008). jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution 25: 1253-1256. Potter, I.C. and G.A. Hyndes (1999). Characteristics of the ichthyofaunas of southwestern Australian estuaries, including comparisons with holarctic estuaries and estuaries elsewhere in temperate Australia: A review. Australian Journal of Ecology 24: 395-421. Prell, W.L., W.H. Hutson, D.R. Williams, A.W.H. Be, K. Geitzenauer and B. Molfino (1980). Surface circulation of the Indian Ocean during the Last Glacial Maximum, approximately 18 000 years BP. Quaternary Research 14: 309-336. 210
References Proebstel, D.S., R.P. Evans, D.K. Shiozawa and R.N. Williams (1993). Preservation of nonfrozen tissue samples from a salmonine fish Brachymystax lenok (Pallas) for DNA analysis. Journal of Ichthyology 32: 9-17. Pruett, C.L. and K. Winker (2008). The effects of sample size on population genetic diversity estimates in song sparrows Melospiza melodia. Journal of Avian Biology 39: 252-256. Qiagen (2006). QIAquick® Spin Handbook. Radchenko, O.A., I.A. Chereshnev, A.V. Petrovskaya, M.V. Nazarkin and E.A. Chegodaeva (2008). Nucleotide sequence variation of the mitochondrial COI gene in several eelpout species of the genus Zoarces (Zoarcidae, Pisces). Russian Journal of Genetics 44(7): 799-807. Rambaut, A. and A. Drummond (2007). Tracer (version 1.4), Available from http://beast.bio.ed.ac.uk/Tracer. Ramos-Onsins, S.E. and J. Rozas (2002). Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution 19: 2092-2100. Ravago, R.G., V.D. Monje and M.A. Juinio-Meñez (2002). Length and sequence variability in mitochondrial control region of the Milkfish, Chanos chanos. Marine Biotechnology 4: 40-50. Raymond, M. and F. Rousset (1995a). GENEPOP: population genetics software for exact tests and ecumenicism. (vers. 4.0). Journal of Heredity 86: 248-249. Raymond, M. and F. Rousset (1995b). An exact test for population differentiation. Evolution 49(6): 1280-1283. Rees, A.J.J., G.K. Yearsley and K. Gowlett-Holmes. (2006). Codes for Australian Aquatic Biota (on-line version). from http://www.marine.csiro.au/caab/. Rice, W.R. (1989). Analyzing tables of statistical tests. Evolution 43: 223-225. Roberts, D.G., C.A. Gray, R.J. West and D.J. Ayre (2010). Marine genetic swamping: hybrids replace an obligately estuarine fish. Molecular Ecology 19: 508-520. Robertson, A.I. (1982). Population dynamics and feeding ecology of juvenile Australian Salmon (Arripis trutta) in Western Port, Victoria. Australian Journal of Marine and Freshwater Research 33: 369-375. Robertson, D.R., F. Karg, R.L. de Moura, B. Victor and G. Bernardi (2006). Mechanisms of speciation and faunal enrichment in Atlantic parrotfishes. Molecular Phylogenetics and Evolution 40: 795-807. Rocha, L.A. and B.W. Bowen (2008). Speciation in coral-reef fishes. Journal of Fish Biology 72: 1101-1121. Rocha, L.A., D.R. Robertson, J. Roman and B.W. Bowen (2005). Ecological speciation in tropical reef fishes. Proceedings of the Royal Society B 272: 573-579. Rockwood, L.L. (2006). Introduction to Population Ecology. Maldan, MA, Blackwell Publishing. Rodriguez, F., J.L. Oliver, A. Marin and J.R. Medina (1990). The general stochastic model of nucleotide substitution. Journal of Theoretical Biology 142(4): 485-501. Rogers, A.R. (1995). Genetic evidence for a Pleistocene population explosion. Evolution 49: 608-615. Rogers, A.R. and H. Harpending (1992). Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution 9: 552-569. Rohfritsch, A. and P. Borsa (2005). Genetic structure of Indian scad mackerel Decapterus russelli: Pleistocene vicariance and secondary contact in the central Indo-West Pacific seas. Heredity 95: 315-322. 211
References Rohlf, F.J., W.S. Chang, R.R. Sokal and J. Kim (1990). Accuracy of estimated phylogenies: Effects of tree topology and evolutionary model. Evolution 44(6): 1671-1684. Rolland, J.-L., F. Bonhomme, F. Lagardère, M. Hassan and B. Guinand (2007). Population structure of the common sole (Solea solea) in the Northeastern Atlantic and the Mediterranean Sea: revisiting the divide with EPIC markers. Marine Biology 151: 327-341. Ronquist, F., J.P. Huelsenbeck and P. van der Mark (2005). MRBAYES 3 Manual. In. Unpublished Report. Florida State University and University of California at San Diego. pp. Ronquist, F., P. van der Mark and J.P. Huelsenbeck (2009). Bayesian phylogenetic analysis using MRBAYES. The Phylogenetic Handbook. A practical approach to phylogenetic analysis and hypothesis testing. P. Lemey, M. Salemi and A.-M. Vandamme. Cambridge, Cambridge University Press: pp. 210-266. Roques, S., J.A. Godoy, J.J. Negro and F. Hiraldo (2004). Organisation and variation of the mitochondrial control region in two vulture species, Gypaetus barbatus and Neophron percnopterus. Journal of Heredity 95(4): 332-337. Rosenberg, N.A. (2007). Statistical tests for taxonomic distinctiveness from observations of monophyly. Evolution 61: 317-323. Ross, K.G., D. Gotzek, M.S. Ascunce and D.D. Shoemaker (2010). Species delimitation: A case study in a problematic ant taxon. Systematic Biology 59(2): 162-184. Rozas, J. and R. Rozas (1999). DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15: 174-175. Ruzzante, D.E., S. Mariani, D. Bekkevold, C. Andre, H. Mosegaard, L.A.W. Clausen, T.G. Dahlgren, W.F. Hutchinson, E.M.C. Hatfield, E. Torstensen, J. Brigham, E.J. Simmonds, L. Laikre, L.C. Larsson, R.J.M. Stet, N. Ryman and G.R. Carvalho (2006a). Biocomplexity in a highly migratory pelagic marine fish, Atlantic herring. Proceedings of the Royal Society B 273: 1459-1464. Ruzzante, D.E., S.J. Walde, V.E. Cussac, M.L. Dalebout, J. Seibert, S. Ortubay and E. Habit (2006b). Phylogeography of the Percichthyidae (Pisces) in Patagonia: roles of orogeny, glaciation, and volcanism. Molecular Ecology 15: 2949-2968. Ryman, N. and S. Palm (2006). POWSIM: a computer program for assessing statistical power when testing for genetic differentiation. Molecular Ecology 6: 600-602. Ryman, N., S. Palm, C. André, G.R. Carvalho, T.G. Dahlgren, P.E. Jorde, L. Laikre, L.C. Larsson, A. Palmé and D.E. Ruzzante (2006). Power for detecting genetic divergence: differences between statistical methods and marker loci. Molecular Ecology 15: 2031-2045. Saccone, C., C. De Giorgi, C. Gissi, G. Pesole and A. Reyes (1999). Evolutionary genomics in Metazoa: the mitochondrial DNA as a model system. Gene 238: 195209. Saitou, N. and M. Nei (1987). The neighbor-joining method: A new method for reconstructing phylogenetic trees. Molecular Biology and Evolution 4: 406-425. Salamon, H., M.R. Segal, A. Ponce de Leon and P.M. Small (1998). Accommodating error analysis in comparison and clustering of molecular fingerprints. Emerging Infectious Diseases 4(2): 159-168. Sale, P.F., R.K. Cowen, B.S. Danilowicz, G.P. Jones, J.P. Kritzer, K.C. Lindeman, S. Planes, N.V.C. Polunin, G.R. Russ, Y.J. Sadovy and R.S. Steneck (2005). Critical 212
References science gaps impede use of no-take fishery reserves. Trends in Ecology and Evolution 20: 74-80. Sánchez-Pérez, R., J. Ballester, F. Dicenta, P. Arús and P. Martínez-Gómez (2006). Comparison of SSR polymorphisms using automated capillary sequencers, and polyacrylamide and agarose gel electrophoresis: Implications for the assessment of genetic diversity and relatedness in almond. Scientia Horticulturae 108: 310-316. Santana, Q.C., M.P.A. Coetzee, E.T. Steenkamp, O.X. Mlonyeni, G.N.A. Hammond, M.J. Wingfield and B.D. Wingfield (2009). Microsatellite discovery by deep sequencing of enriched genomic libraries. Biotechniques 46: 217-223. Scandol, J., K. Rowling and K. Graham (Eds.) (2008). Status of Fisheries Resources in NSW 2006/07, NSW Department of Primary Industries, Cronulla, NSW: 334 pp. Schmidt, H.A. (2009). Testing tree topologies. The Phylogenetic Handbook. A practical approach to phylogenetic analysis and hypothesis testing. P. Lemey, M. Salemi and A.-M. Vandamme. Cambridge, Cambridge University Press: pp. 381-403. Schmidt, H.A., K. Strimmer, M. Vingron and A. von Haeseler (2002). TREE-PUZZLE: maximum likelihood phylogenetic analysis using quartets and parallel computing. Bioinformatics 18: 502-504. Schmidt, H.A. and A. von Haeseler (2009). Phylogenetic inference using maximum likelihood methods. The Phylogenetic Handbook. A practical approach to phylogenetic analysis and hypothesis testing. P. Lemey, M. Salemi and A.-M. Vandamme. Cambridge, Cambridge University Press: pp. 181-209. Sefc, K.M., R.B. Payne and M.D. Sorenson (2003). Microsatellite amplification from museum feather samples: effects of fragment size and template concentration on genotyping errors. The Auk 120(4): 982-989. Selander, R.K. (1976). Genic variation in natural populations. Molecular Evolution. F.J. Ayala. Sunderland, Massachusetts, Sinauer: pp. 21-45. Shaklee, J.B. and K.P. Currens (2003). Genetic stock identification and risk assessment. Population Genetics: Principles and Applications for Fisheries Scientists. E.M. Hallerman. Bethesda, American Fisheries Society: pp. 291-328. Shimodaira, H. (2002). An approximately unbiased test of phylogenetic tree selection. Systematic Biology 51: 492-508. Shimodaira, H. and M. Hasegawa (1999). Multiple comparisons of log-likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16: 11141116. Shimodaira, H. and M. Hasegawa (2001). CONSEL: for assessing the confidence of phylogenetic tree selection. Bioinformatics 17: 1246-1247. Shinde, D., Y. Lai, F. Sun and N. Arnheim (2003). Taq DNA polymerase slippage mutation rates measured by PCR and quasi-likelihood analysis: (CA/GT)n and (A/T)n microsatellites. Nucleic Acids Research 31: 974-980. Siegel, D., B.P. Kinlan, B. Gaylord and S.D. Gaines (2003). Lagrangian descriptions of marine larval dispersion. Marine Ecology Progress Series 260: 83-96. Siegel, S. and J.N.J. Castellan (1988). Nonparametric Statistics for the Behavioral Sciences. New York. (Table T. Critical values for Kendall's coefficient of concordance W. p.365), McGraw-Hill Book Company. Sites, J.W.J. and J.C. Marshall (2004). Operational criteria for delimiting species. Annual Review of Ecology and Systematics 35: 199-227. Slatkin, M. and R.R. Hudson (1991). Pairwise comparisons of mitochondrial DNA sequences in stable and exponentially growing populations. Genetics 129: 555-562. 213
References Slowinski, J.B. (2001). Molecular polytomies. Molecular Phylogenetics and Evolution 12(1): 114-120. Smith, J.R., J.D. Carpten, M.J. Brownstein, S. Ghosh, V.L. Magnuson, D.A. Gilbert, J.M. Trent and F.S. Collins (1995). Approach to genotyping errors caused by nontemplated nucleotide addition caused by Taq DNA polymerase. Genome Research 5: 312-317. Smith, K. and J. Brown (2007a). Australian Salmon Managed Fisheries Status Report. State of the Fisheries Report 2006/07. W.J. Fletcher and K. Santoro. Western Australia, Department of Fisheries: 229-232. Smith, K. and J. Brown (2007b). Australian Herring Managed Fisheries Status Report. State of the Fisheries Report 2006/07. W.J. Fletcher and K. Santoro. Western Australia, Department of Fisheries: 233-237. Sotka, E.E., J.A. Hempelmann and C.H. Biermann (2005). Genetic evidence of postglacial population expansion in Puget Sound Rockfish (Sebastes emphaeus). Marine Biotechnology 7(3): 223-230. Sourdis, J. and M. Nei (1988). Relative efficiencies of the maximum parsimony and distance-matrix methods in obtaining the correct phylogenetic tree. Molecular Biology and Evolution 5: 298-311. Spencer, H.G., J.M. Waters and T.E. Eichhorst (2007). Taxonomy and nomenclature of black nerites (Gastropoda: Neritimorpha: Nerita) from the South Pacific. Invertebrate Systematics 21: 229-237. Stanley, C.A. (1978). Area of distribution, movements, age composition and mortality rates of the Australian Salmon population in Tasmania, Victoria and New South Wales. Australian Journal of Marine and Freshwater Research 29: 417-433. Stanley, C.A. (1980). Lee's Phenomenon in the western subspecies of the Australian Salmon, Arripis trutta esper. Australian Journal of Marine and Freshwater Research 31: 13-19. Stanley, C.A. (1986). Tagging experiments on Australian Salmon (Arripis trutta esper): Recapture data for South Australian releases, 1952 to 1976. Unpublished Report. CSIRO Marine Laboratories, Hobart, Tasmania. pp. Stanley, C.A. (1988a). Tagging experiments on Australian Salmon (Arripis trutta): Recapture data for Tasmanian releases, 1949 to 1964. Unpublished Report. CSIRO Marine Laboratories, Hobart, Tasmania. pp. Stanley, C.A. (1988b). Tagging experiments on Australian Salmon (Arripis trutta esper): Recapture data for Western Australian releases, 1945 to 1976. Unpublished Report. CSIRO Marine Laboratories, Hobart, Tasmania. pp. Stanley, C.A. (1988c). Tagging experiments on Australian Salmon (Arripis trutta): Recapture data for New South Wales releases, 1939 to 1965, and Victorian releases, 1939 to 1964. Unpublished Report. CSIRO Marine Laboratories, Hobart, Tasmania. pp. Stanley, C.A. and W.B. Malcolm (1977). Reproductive cycles in the eastern subspecies of the Australian Salmon, Arripis trutta marginata (Cuvier & Valenciennes). Australian Journal of Marine and Freshwater Research 28: 287-301. Stewart, J., J.M. Hughes, J. McAllister, J.M. Lyle and C.M. MacDonald (2011). Australian salmon (Arripis trutta): Population structure, reproduction, diet and composition of commercial and recreational catches. Unpublished Report. Investment and Industry NSW, Cronulla, Cronulla. 259 pp. pp. Stres, B. (2006). The first decade of terminal restriction fragment length polymorphism (T-RFLP) in microbial ecology. Acta agriculturae Slovenica 88: 65-73. 214
References Strimmer, K. and A. von Haeseler (1996). Quartet-puzzling: a quartet maximumlikelihood method for reconstructing tree topologies. Molecular Biology and Evolution 13: 964-969. Swofford, D.L. (2005). PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sunderland, Massachusetts, Sinauer Associates. Swofford, D.L., G.J. Olsen, P.J. Waddell and D.M. Hillis (1996). Phylogenetic inference. Molecular Systematics. 2nd Edition. D.M. Hillis, C. Moritz and B.K. Mable. Sunderland, Massachusetts, Sinauer Associates, Inc.: pp. 407-514. Swofford, D.L. and J.C. Sullivan (2009). Phylogeny inference based on parsimony and other methods using PAUP*. The Phylogenetic Handbook. A practical approach to phylogenetic analysis and hypothesis testing. P. Lemey, M. Salemi and A.-M. Vandamme. Cambridge, Cambridge University Press: pp. 267-312. Symonds, V.V. and A.M. Lloyd (2004). A simple and inexpensive method for producing fluorescently labelled size standard. Molecular Ecology Notes 4: 768771. Szpiech, Z.A., M. Jakobsson and N. Rosenberg (2008). ADZE: A rarefaction approach for counting alleles private to combinations of populations. Bioinformatics 24(21): 2498-2504. Taberlet, P., S. Griffin, B. Goossens, S. Questiau, V. Manceau, N. Escaravage, L.P. Waits and J. Bouvet (1996). Reliable genotyping of samples with very low DNA quantities using PCR. Nucleic Acids Research 24: 3189-3194. Tajima, F. (1989). Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585-595. Takagi, A.P., S. Ishikawa, T. Nao, S. Hort, M. Nakatani, M. Nishida and H. Kurokura (2006). Tandem repeat sequences in control region of bronze featherback Notopterus notopterus mitochondrial DNA. Fisheries Science 72: 1319-1321. Takezaki, N. and M. Nei (1996). Genetic distances and reconstruction of phylogenetic trees from microsatellite DNA. Genetics 144: 389-399. Tamura, K., J. Dudley, M. Nei and S. Kumar (2007). MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Molecular Biology and Evolution 24: 1596-1599. Tamura, K. and M. Nei (1993). Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution 10: 512-526. Tavare, S. (1986). Some probabilistic and statisical problems on the analysis of DNA sequences. Lectures in Maths and Life Sciences 17: 57-86. Tavares, E. and A. Baker (2008). Single mitochondrial gene barcodes reliably identify sister-species in diverse clades of birds. BMC Evolutionary Biology 8: 81-94. Templeton, A.R. (1998). Nested clade analysis of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7: 381-397. Templeton, A.R. (2004). Using haplotype trees for phylogeographic and species inference in fish populations. Environmental Biology of Fishes 69: 7-20. Templeton, A.R., K.A. Crandall and C.F. Sing (1992). A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 132: 619633. Templeton, A.R., E. Routman and C.A. Phillips (1995). Seperating population structure from population history: a cladistic analysis of the geographical distribution of 215
References mitochondrial DNA haplotypes in the tiger salamander, Ambystoma tigrinum. Genetics 140: 767-782. Teske, P.R., M.I. Cherry and C.A. Matthee (2004). The evolutionary history of seahorses (Syngnathidae: Hippocampus): molecular data suggest a West Pacific origin and two invasions of the Atlantic Ocean. Molecular Phylogenetics and Evolution 30: 273-286. Teske, P.R., H. Hamilton, C.A. Matthee and N.P. Barker (2007). Signatures of seaway closures and founder dispersal in the phylogeny of a circumglobally distributed seahorse lineage. BMC Evolutionary Biology 7: 138-156. Theisen, T.C., B.W. Bowen, W. Lanier and J.D. Baldwin (2008). High connectivity on a global scale in the pelagic wahoo, Acanthocybium solandri (tuna family Scombridae). Molecular Ecology 17: 4233-4247. Thompson, J.M. (1957). The food of Western Australian estuarine fish. Fisheries Bulletin (Western Australian Fisheries and Fauna Department) 7: 1-13. Tindall, K.R. and T.A. Kunkel (1988). Fidelity of DNA synthesis by the Thermus aquaticus DNA polymerase. Biochemistry 27(16): 6008-6013. Toonen, R.J., K.R. Andrews, I.B. Baums, C.E. Bird, G.T. Concepcion, T.S. Daly-Engel, J.A. Eble, A. Faucci, M.R. Gaither, M. Iacchei, J.B. Puritz, J.K. Schultz, D.J. Skillings, M.A. Timmers and B.W. Bowen (2011). Defining boundaries for ecosystem-based management: A multispecies case study of marine connectivity across the Hawaiian Archipelago. Journal of Marine Biology 2011(Article ID 460173): 13 pp. TPC (2009). State of the Environment Report: Tasmania 2009. Unpublished Report. Tasmanian Planning Commission, Hobart. pp. Tregonning, J.E., R.C.J. Lenanton and D.A. Hancock (1995). Biological and fisheries synopsis of Australian herring, Arripis georgianus (Valenciennes, 1831), (Arripididae: Percoidei). Unpublished Report. Fisheries Department of Western Australia. pp. Uzzell, T. and K.W. Corbin (1971). Fitting discrete probability distributions to evolutionary events. Science 172: 1089-1096. Valenciennes, A. (1831). Histoire naturelle des poissons. Tome septième. Livre septième. Des Squamipennes. Livre huitième. Des poissons à pharyngiens labyrinthiformes. Histoire naturelle des poissons. G. Cuvier and A. Valenciennes. 7: 1-531. Vekemans, X., T. Beauwens, M. Lemaire and I. Roldán-Ruiz (2002). Data from amplified fragment length polymorphism (AFLP) markers show indication of size homoplasy and of a relationship between degree of homoplasy and fragment size. Molecular Ecology 11: 139-151. Villablanca, F.X., G.K. Roderick and S.R. Palumbi (1998). Invasion genetics of the Mediterranean fruit fly: variation in multiple nuclear introns. Molecular Ecology 7: 547-560. Waddell, J. (2011). Indigenous cultural uses and values of Eastern Australian Salmon in southern NSW. Australian salmon (Arripis trutta): Population structure, reproduction, diet and composition of commercial and recreational catches. J. Stewart et al. Cronulla, Industry & Investment NSW - Fisheries Final Report Series No. 129: pp. 238-252. Waddell, V., M.C. Milinkovitch, M. Berube and M.J. Stanhope (2000). Molecular phylogenetic examination of the Delphinoidea trichotomy: congruent evidence from three nuclear loci indicates that Porpoises (Phocoenidae) share a more recent 216
References common ancestry with White Whales (Monodontidae) than they do with true Dolphins (Delphinidae). Molecular Phylogenetics and Evolution 15: 314-318. Wägele, J.W., H.H. Letsch, A. Klussmann-Kolb, C. Mayer, B. Misof and H. Wägele (2009). Phylogenetic support values are not necessarily informative: the case of the Serialia hypothesis (a mollusk phylogeny). Frontiers in Zoology 6: 12-26. Wakeley, J. and J. Hey (1997). Estimating ancestral population parameters. Genetics 145: 847-855. Walsh, H.E., M.G. Kidd, T. Moum and V.L. Friesen (1999). Polytomies and the power of phylogenetic inference. Evolution 53: 932-937. Wang, M., X. Zhang, T. Yang, Z. Han, T. Yanagimoto and T. Gao (2008). Genetic diversity in the mtDNA control region and population structure in the Sardinella zunasi Bleeker. African Journal of Biotechnology 7(24): 4384-4392. Waples, R.S. (1987). A multispecies approach to the analysis of gene flow in marine shore fishes. Evolution 41(2): 385-400. Ward, R. (2009). DNA barcode divergence among species and genera of birds and fishes. Molecular Ecology Resources 9(4): 1077-1085. Ward, R.D., M. Woodwark and D.O.F. Skibinski (1994). A comparison of genetic diversity levels in marine, freshwater, and anadromous fishes. Journal of Fish Biology 44: 213-232. Ward, R.D., T.S. Zemlak, B.H. Innes, P.R. Last and P.D.N. Hebert (2005). DNA barcoding Australia's fish species. Philosophical Transactions of the Royal Society B: Biological Sciences 360: 1847-1857. Wares, J.P. and C.W. Cunningham (2001). Phylogeography and historical ecology of the North Atlantic intertidal. Evolution 55: 2455-2469. Waters, J.M., T.M. King, P.M. O’Loughlin and H.G. Spencer (2005). Phylogeographical disjunction in abundant high-dispersal littoral gastropods. Molecular Ecology 14: 2789-2802. Waters, J.M., P.M. O'Loughlin and M.S. Roy (2004). Cladogenesis in a starfish species complex from southern Australia: evidence for vicariant speciation? Molecular Phylogenetics and Evolution 32: 236-245. Waters, J.M. and M.S. Roy (2003). Marine biogeography of southern Australia: phylogeographical structure in a temperate sea-star. Journal of Biogeography 30: 1787–1796. Watterson, G.A. (1984). Allele frequencies after a bottleneck. Theoretical Population Biology 26: 387-407. Webb, B.F. (1973a). Fish populations of the Avon-Heathcote estuary: 2. Breeding and gonad maturity. New Zealand Journal of Marine and Freshwater Research 7(1 & 2): 45-66. Webb, B.F. (1973b). Fish Populations of the Avon-Heathcote estuary: 3. Gut contents. New Zealand Journal of Marine and Freshwater Research 7(3): 223-234. Weir, B.S. and C.C. Cockerham (1984). Estimating F-statistics for the analysis of population structure. Evolution 38(6): 1358-1370. Weir, J.T. and D. Schluter (2008). Calibrating the avian molecular clock. Molecular Ecology 17: 2321-2328. Wells, P. and H. Okada (1996). Holocene and Pleistocene glacial palaeoceanography off southeastern Australia, based on foraminifers and nannofossils in Vema cored hole V18-222. Australian Journal of Earth Sciences 43(5): 509-523.
217
References Wells, P.E. and G.M. Wells (1994). Large-scale reorganisation in ocean circulation off Western Australia during the Late Pleistocene. Marine Micropaleontology 24: 157186. Whitfield, J.B. and P.J. Lockhart (2007). Deciphering ancient rapid radiations. Trends in Ecology and Evolution 22: 258-265. Whitley, G. (1932). Marine zoogeographical regions of Australia. Australian Naturalist 8: 166-167. Willerslev, E., M.T.P. Gilbert, J. Binladen, S.Y.W. Ho, P.F. Campos, A. Ratan, L.P. Tomsho, R.R. da Fonseca, A. Sher, T.V. Kuznetsova, M. Nowak-Kemp, T.L. Roth, W. Miller and S.C. Schuster (2009). Analysis of complete mitochondrial genomes from extinct and extant rhinoceroses reveals lack of phylogenetic resolution. BMC Evolutionary Biology 9: 95 doi:10.1186/1471-2148-9-95. Williams, S. (2007). Salmon that ate NSW. The Daily Telegraph. Sydney. Williams, S.T., N. Knowlton, L.A. Weigt and J.A. Jara (2001). Evidence for three major clades within the snapping shrimp genus Alpheus inferred from nuclear and mitochondrial gene sequence data. Molecular Phylogenetics and Evolution 20: 375389. Wilson, B.R. and G.R. Allen (1987). Major components and distribution of marine fauna. Fauna of Australia. G.R. Dyne and D.W. Walton. Canberra, Australian Government Publishing Service: pp. 43-68. Wirth, T. and L. Bernatchez (2001). Genetic evidence against panmixia in the European eel. Nature 409: 1037-1040. Wolf, J.B.W., D. Tautz and F. Trillmich (2007). Galápagos and Californian sea lions are separate species: Genetic analysis of the genus Zalophus and its implications for conservation management. Frontiers in Zoology 4: 20-32. Womersley, H.B.S. (1981). Biogeography of Australian marine macroalgae. Marine Botany: an Australian perspective. W.N. Clayton and R.J. King. Melbourne, Longman Cheshire: pp. 292–307. Woodroffe, C.D., D.M. Kennedy, B.P. Brooke and M.E. Dickson (2006). Geomorphological evolution of Lord Howe Island and carbonate production at the latitudinal limit reef growth. Journal of Coastal Research 22(1): 188-201. Xia, X. (2009). Assessing substitution saturation with DAMBE. The Phylogenetic Handbook. A practical approach to phylogenetic analysis and hypothesis testing. P. Lemey, M. Salemi and A.-M. Vandamme. Cambridge, Cambridge University Press: pp. 615-630. Xia, X., Z. Xie, M. Salemi, L. Chen and Y. Wang (2003). An index of substitution saturation and its application. Molecular Phylogenetics and Evolution 26: 1-7. York, K.L., M.J. Blacket and B.R. Appleton (2008). The Bassian Isthmus and the major ocean currents of southeast Australia influence the phylogeography and population structure of a southern Australian intertidal barnacle Catomerus polymerus (Darwin). Molecular Ecology 17: 1948-1961. Zink, R.M. and G.F. Barrowclough (2008). Mitochondrial DNA under siege in avian phylogeography. Molecular Ecology 17: 2107-2121.
218
Appendix A.1 Appendix A.1 Characteristics and amplification of mtDNA control region in Arripis.
As stated in Chapter 2 (Section 2.4.3), attempts to find reliable primers for the PCR amplification of the control region in the mtDNA of Arripis were unsuccessful despite the fact that a range of primer pairs and approaches was used. An overview of the characteristics and amplification of the mtDNA control region in Arripis is provided below.
Nucleotide sequence information was obtained from the control region of some individuals of A. trutta and A. truttaceus but not A. georgianus (using combinations of primer pairs Atr15614L, Atr15621L with Atr16229H, Atr16295H, Atr16311H; Table A.1). However, less than 60% of PCR amplification attempts resulted in interpretable sequences and no single set of primers was successful in all of those individuals from which the control region was amplified. Furthermore, despite extensive optimisation, double bands and smears (in the post-PCR gel) could not be eliminated in any of the samples that did amplify. No primers amplified any part of the control region sequence in A. georgianus. No attempt was made to amplify the control region in A. xylabion.
As implied above, mtDNA sequences of inconsistent length and position were generated for A. trutta and A. truttaceus. In addition, the presence a 55 bp variable number of tandem repeats (VNTR) was discovered at the 5′ end of the d-loop region of the control region (VNTR motif: 5′-ATG TTG ACC ATA CAT ATA TGC TTT AAA TAC ATA TAT GTA TAA TCA CCA TAA ATT T-3′). In A. trutta, between five and eight copies of the VNTR were detected among 13 sequences, and the last repeat was always 49 bp long (the last 6 bp had been deleted; an imperfect repeat at the 3’ end is a well known 219
Appendix A.1 phenomenon in VNTR – see Fumigalli et al. 1996). In A. truttaceus, between six and nine copies of the VNTR were detected among eight sequences, and all repeats were of equal length (i.e. no deletions). As is typical for VNTR (Lee et al. 1995; Lunt et al. 1998), the repeat sequence was AT rich (78%).
The incidence of VNTR in the mtDNA of vertebrates, including teleosts, appears to be quite common (e.g. Lee et al. 1995; Chen et al. 1998; Nesbø et al. 1998; Kong et al. 2007), and is usually attributed to slipped-strand replication derived from repair mechanisms and secondary structures that induce misalignment prior to elongation and replication of the d-loop strand (Lee et al. 1995; Fumigalli et al. 1996). Alignment of the repeat motif of Arripis with its reverse compliment revealed that most of the sequence was complementary, suggesting slipped-strand replication might be the source of VNTR in this genus (Gemmell et al. 1996).
VNTR have been useful markers in a range of vertebrates, using both sequence based or allele-based (repeat number) analyses (Broughton and Dowling 1997; Chen et al. 1998; Lunt et al. 1998; Nesbø et al. 1998; Ravago et al. 2002; Roques et al. 2004; Ketmaier and Bernardini 2005; Takagi et al. 2006; Kong et al. 2007). However, several hurdles still remain. For example, the presence of VNTR in the control (d-loop) region can cause problems during enzyme reading in the PCR reaction making PCR failure a common problem even when trying to amplify larger sequences of which the VNTR is only a small part (e.g. Lunt et al. 1998). Secondly, while the number of repeat copies might be scorable (as alleles) by counting bands on an electrophoretic gel, multiple heterogeneous motifs provide uninterpretable sequences downstream of the VNTR (due to multiple peaks on a chromatogram); this was common in the Arripis chromatograms. 220
Appendix A.1 A third serious problem arises from both unreliable PCR and heteroplasmy; repeated PCR-amplifications of the same individual may result in different estimates of the number of repeat motifs (Lunt et al. 1998), as was the case in the assays of Arripis.
Ultimately, in this study, the control region, and its VNTR component in particular, were rejected as a useful genetic marker for the species of Arripis for the following reasons. Firstly, as described above, at least two technical hurdles regarding the use of VNTR are yet to be resolved (slippage and heteroplasmy) and both appear likely to be affecting the Arripis sequence. Secondly, despite targeting priming sites inside and outside of the control region (Table A.1), amplification of PCR product with interpretable sequence in A. trutta and A. truttaceus was only possible in a few individuals, and not at all possible in A. georgianus. To avoid these problems, some workers have avoided VNTR by targeting primer sites within Conserved Sequence Blocks (CSB), especially CSB-D (e.g. Lee et al. 1995), however, this was also unsuccessful in Arripis (see below).
Many primer combinations were used in an attempt to PCR-amplify the mtDNA control region. Three different approaches were used to find reliable primers, namely: universal primers, customised primers and primer walking.
The first approach applied universal fish primers (Kocher et al. 1989; Meyer et al. 1990; Table A.1) that have been used to amplify partial-fragments of the d-loop of the control region in a range of other perciform species (e.g. Bearham 2004), but without success in Arripis.
221
Appendix A.1 In a second approach, attempts were made to develop customised primers to amplify the control region in two ways. Firstly, a complete unpublished mtDNA genome sequence for a single individual of A. trutta (provided by M. Miya, Natural History Museum, Chiba, Japan) was used to design a range of A. trutta specific primers in: (i) the tRNA genes at each end of the control region; (ii) the conserved region (CSB) in the middle of the control region; (iii) the position of universal d-loop fish primers; (iv) cytochrome b; and (v) 16S rRNA (Table A.1). As described above, some PCR-amplifications were achieved, but were not reliable and were complicated by the discovery of the VNTR. Secondly, using the BLAST function of GenBank (Benson et al. 2007), species for which parts of the mtDNA were similar to Arripis were identified.
Sequence
alignments involving these identified species were used to design primers based in the conserved parts of the genes that flank the control region (tRNA Threonine, tRNA Proline, tRNA Phenylalanine; Table A.1), although none successfully amplified any PCR product.
In the third approach, alignments, as described above, were used to design primers in the cytochrome b and 12S rRNA genes and fill in the ~ 5 kb gap between (spanning the control region) using ‘primer walking’ (Gromek and Kaczorowski 2005; Table A.1). This failed to amplify any PCR product, even when high-fidelity long-range polymerase was used (Phusion™ Flash; Finnzymes Oy).
222
Appendix A.1 Table A.1
MtDNA primers tested for PCR-amplification of the control region in
this study. The region/gene in which the primer was situated is indicated. Sequences use IUPAC codes for nucleotides. Primer Name
Region
Strand
Sequence (5′ – 3′)
L15926
tRNAProline
forward
AACTCTCACCCCTAGCTCCCAAAG
H16498
control region
reverse
CCTGAAGTAGGAACCAGATG
Universal Primers
Primers developed from indirect sequence alignment tRNA-Thr F tRNA-Pro F tRNA-Phe R
tRNAThreonine
forward
TAARCCGRACGYCGGAGGTT
Proline
forward
TTTARTTTAGAATYCTRGCTTTGG
Phenylalanine
reverse
TCTTAACAKCTTCAGTGTTATGCT
tRNA tRNA
Primers developed from Arripis trutta sequence Atr15806L
tRNAProline
forward
ATTCTACAACATTAAACTATTCCTT
Atr15828L
Proline
tRNA
forward
CTTGCGAATAACTATGTTGACA
Atr16229H
control region
reverse
GATAGGTTTCATTAGGTTACACTT
Atr16295H
control region
Atr17270H-Phe
reverse
CCTGATGGTCGGTTCTTATTAC
Phenylalanine
reverse
GCTCATCTCAACATCTTCAGTG
Proline
tRNA
Atr15778L
tRNA
forward
AACTCTTACCCTTAGCTCCCA
Atr16311H
control region
reverse
CATTAAGTAATAACCTGATG
Atr16495H
control region
reverse
TTGGAGTGAACGCTAGGCAT
Atr16746H
control region
reverse
AGTTTTGGGGGGGTAGG
Atr15065L-Cytb
cytochrome b
forward
TACTCTCCCGATTCTTTGCCTTCC
Atr01250L-16S
16S rRNA
forward
GACAGAAAAAGAACTATTGGAGCT
Atr01612H-16S
16S rRNA
Atr15737L
reverse
CTTTGACTAACTCTCCTGGTTTCTC
Threonine
forward
CGTTCAAACCTTCCCTCA
Threonine
tRNA
Ageo15737L
tRNA
forward
CGGTGAATACCTGCATACTAA
Atr15614L
control region
forward
CCTCTACTTCTCGTTATTCCTCATCCT
Atr15621L
control region
forward
TTCTCGTTATTCCTCATCCTC
Atr16383H
control region
reverse
GGAACCTATGCCAGGAATA
Atr16739H
control region
reverse
GGGGGGGTAGGGGGGGTA
Primers developed for primer walking Arripis Cytb
cytochrome b
forward
TCCGTATGATGAAACTTC
Atrutt Cytb
cytochrome b
forward
CGAGAACAAGGTTTTAGGATGA
Arripis 12S
12S rRNA
reverse
TCTCGGTGTAAGGGAAA
Ageorg 12S
12S rRNA
reverse
CTTCTCGGTGTAAGGGTAATGC
223
Appendix A.2 EPIC-PCR primers tested in this study. Genes in bold were polymorphic in at least one species and subsequently assayed. Genes marked with an asterisk (*) amplified, but were monomorphic within each species. Sequences use IUPAC codes for nucleotides.
Gene
Strand
Primer Sequence (5′ – 3′)
Reference/s
PmActI-F PmActI-R
forward reverse
CCATACCTTCTACAATGAGCTCCG GACCAGAGGCGTAGAGGAGAGC
1, 2, 3, 4
Actin (intron 2)
Act2
Act-2-F Act-2-R
forward reverse
GCATAACCCTCGTAGATGGGCAC ATCTGGCACCACACCTTCTACAA
2, 3, 5
Aldolase B (intron 1) *
AldoB1
AldoB1-1F AldoB1-2R
forward reverse
GCTCCAGGAAAGGGAATCCTGGC CTCGTGGAAGAAGATGATCCCGCC
6, 7
Aldolase B (intron 2)
AldoB2
AldoB2F Aldexon3R
forward reverse
TCAGGGCATTGTCGTCGG CTGTTCCATTTAGACCAGC
2, 6, 7
Aldolase B (intron 4)
AldoB4
Aldo5F Aldo3.1R
forward reverse
GCCAGATATGCCAGCATCTGCC GGGTTCCATCAGGCAGGATCTCTGGC
6, 7
Aldolase C (intron 1)
AldoC1
AldoC1F AldoC2R
forward reverse
CCTGGCTGCGGACGAGTCTGTGGG GGCGGTACTGTCTGCGGTTCTCC
2, 5, 6
Alpha amylase (intron 1) *
Am2B-1
Am2B1F Am2B2R
forward reverse
CCTTCATCTTCCAGGAGGTAC TTCACCTCCCAGATCAATAAC
6
Alpha amylase (intron 2)
Am2B-2
Am2B2F Am2B3R
forward reverse
GGCGATAAGTTGTCTTACACC AGCCCTCTCCCCAGTTCCTGC
6, 8
β-globin (intron 1)
β-globin
β-1 β-2
forward reverse
GTTGGTGGTGAGGCCCTGGGCAG CCTGAAGTTCTCAGGATCCACATGCA
9
Calmodulin (intron 3)
CaM3
Cam-3-F Cam-3-R
forward reverse
TGACGGAGCTCTGCAGCACTGAC GTGAGGAGGAGCTCCGTGAGGC
5
Appendix A.2
Primer Name
Actin (intron 1)
Intron Name Act1
Appendix A.2 continued.
Gene
Strand
Primer Sequence (5′ – 3′)
Reference/s
CalMex4F CalMex5R
forward reverse
CTGACCATGATGGCCAGAAA GTTAGCTTCTCCCCCAGGTT
3, 10, 11, 12
Creatin kinase (intron 7)
CK7
CK7F CK8R
forward reverse
AAGAGGGTCTTTGACAGGTTCTGC TTCTCCTGGATCAGACGCTCCACC
6, 7, 13
Elongation factor -1α (intron 6)
EF1α6
EF16cF EF17bR
forward reverse
GGAAACKTGGCTGGAGAC AATGGCAGCATCTCCAGACT
11
Elongation factor (intron 2)
EF2
EF2-F EF2-R
forward reverse
GTCTTCGTGCAGGAGAAACTCG GAGGGCGAGGTTCTCGTCACTCCAGC
1
Glucose-6-phosphate dehydrogenase (intron 4)
G6PD
G6PDex4F G6PDex5R
forward reverse
GAGCAGACGTATTTTGTGGG GCCAGGTAGAAGAGGCGGTT
14
Growth hormone (intron 1)
GH1
SG1 FIS
forward reverse
AGAACCTGAACCAGACATGG AGGTGTTGAACTCTGCTGAC
15
Growth hormone (intron 2)
GH2
GH2F GH3R
forward reverse
AGCGTTTCTCCATTGCCGTCAGC TCTTGTTGAGTTGACGCTGGTCC
6, 7
Growth hormone (intron 5)
GH5
GH5F GH6R
forward reverse
AGCCAATCAGGACGGAGC TGCCACTGTCAGATAAGTCTCC
6, 7, 8
Gonadotropin-releasing hormone 3 (intron 1) *
GnRH3-1
GnRH1F GnRH1R
forward reverse
AATGCACCACATGCTAACAAGGC CGCACCATCACTCTGCTGTTCGC
6, 7, 8
Gonadotropin-releasing hormone 3 (intron 2) *
GnRH3-2
GnRH2F GnRH2R
forward reverse
AGAAGTGTGGGAGAGCTAGAGGC AGAGACACCACTTCTCCTGTACCC
6, 7, 16
Appendix A.2
Primer Name
Calmodulin (intron 4)
Intron Name CaM4
Appendix A.2 continued.
Gene Glyceraldehyde-3-phosphate dehydrogenase (intron 2) *
Intron Name Gpd2
Primer Name
Strand
Primer Sequence (5′ – 3′)
Reference/s
GPD2F GPD3R
forward reverse
GCCATCAATGACCCCTTCATCG TTGACCTCACCCTTGAAGCGGCCG
6, 7, 8
Major histocompatibility complex class II antigen (intron 1)
MhcII
Mhc1F Mhc2R
forward reverse
ACTCTAATCTGGAGTACATGC CAGGAGATCTTCTCTCCAGCC
6
Myosin light chain B (intron 2) * †
Mlc-2-c
Mlc-2-F Mlc-2-R
forward reverse
CTCGGGGTCCAGGACCTTGAA CTTCACCGTCGACCTCACCATG
2, 5
Opsin (intron 1)
Ops1
Ops-1-F Ops-1-R
forward reverse
GCTCATGGGCCTGCAGACCACAA CCTGCTCAACCTGGCCATGGC
5, 8, 10
S7 ribosomal protein (intron 1) *
S71
S7RPEX1F S7RPEX2R
forward reverse
TGGCCTCTTCCTTGGCCGTC AACTCGTCTGGCTTTTCGCC
8, 17, 18, 19, 20, 21, 22,23
S7 ribosomal protein (intron 2)
S72
S7RPEX2F S7RPEX3R
forward reverse
AGCGCCAAAATAGTGAAGCC GCCTTCAGGTCAGAGTTCAT
17, 18, 21, 22, 23
Alpha tropomyosin (intron 1)
Tr1
Tr1F Tr1R
forward reverse
AGGGAACAGAGGATGAGCTGGAC TCTCAGCTTCCTCCAGCTTGGTG
6
1.Bierne et al. 2000; 2. Berrebi et al. 2005; 3. Berrebi et al. 2006; 4. Villablanca et al. 1998; 5. Atarhouch et al. 2003; 6. Hassan et al. 2002; 7. Fauvelot et al. 2007; 8. Hubert et al. 2006; 9. Lessa 1992; 10. Atarhouch et al. 2007; 11. Moyer et al. 2005; 12. Chow and Takeyama 2000; 13. Palumbi and Baker 1994; 14. Chow and
21. Lavoué et al. 2003; 22. Burridge et al. 2006; 23. Cooke and Beheregaray 2007.
†
The primers for Mlc-2-c amplified two putative loci.
Appendix A.2
Nakadate 2004; 15. Almuly et al. 2000; 16. Ruzzante et al. 2006b; 17. Chow and Hazama 1998; 18. Chow et al. 2001; 19. Morrison et al. 2006; 20. Teske et al. 2004;
Appendix A.3 Appendix A.3 Technical issues regarding the use of fragment length polymorphism at intron loci.
The present study used nDNA markers in the form of fragment length polymorphism at intron loci. This type of nDNA marker was used because it provided many of the advantageous qualities of microsatellite markers (e.g. locus-specific, co-dominance, relatively high amounts of fragment length polymorphism; e.g. Li et al. 2002), but which could be amplified using universal primers (and thereby avoid the time and expense of developing customised microsatellite primers; see Chapter 2). Since this is one of few studies to use such markers to address population and evolutionary questions, and the first to use automated electrophoretic scoring methods with this type of marker, it is worthwhile to briefly consider some of the technical issues associated with the production of the intron data.
As discussed in Chapter 2, there were some ambiguities associated with determining the sizes of the alleles, particularly larger ones, at the intron loci. These ambiguities were manifested in two ways. Firstly, the raw estimates of the sizes of one or both alleles at a locus in repeat assays of the same individuals were sometimes different (by up to 1.5 bp; see Section 2.5.4). Secondly, the frequencies of raw estimates of the sizes of an allele at a locus tended to follow a normal distribution (Figures A.1, A.2), which was interpreted as reflecting minor scoring errors of a single allele. It is likely that aspects of both the PCR and post-PCR assays contributed to these ambiguities, as considered below.
227
Appendix A.3
45 40
Frequency
35 30 25 20 15 10 5 179
178
177
0
Fragment Length
Figure A.1
Example of the frequency distribution of the raw estimates of the
fragment sizes around the 178 bp allele of locus AldoB4 in all assayed Arripis trutta.
4.5
Frequency
4 3.5 3 2.5 2 1.5 1 0.5 903
902
901
900
899
898
897
0
Fragment Length
Figure A.2
Example of the frequency distribution of the raw estimates of the
fragment sizes around the 900 bp allele of locus S72 in all assayed Arripis trutta. The modal integer is 900.
228
Appendix A.3 PCR Assays Due to limitations in the replication fidelity of the Taq polymerase (e.g. Tindall and Kunkel 1988), there may have been small differences in the sequence compositions in the amplicons of identical alleles in the present study (e.g. Taberlet et al. 1996; Davidson et al. 2003; Shinde et al. 2003). In absolute terms, the potential for these errors is obviously greater in longer fragments. This in turn, could have affected the electrophoretic mobility of the amplicons (see below).
Post-PCR Assays Intron ‘alleles’ of the same length (number of base pairs) do not necessarily have identical sequence compositions and even a slight difference in composition can result in minor disparities in electrophoretic mobility (Anon. 2004). This effect is expected to be greater in longer fragments because the absolute number of mutations (and hence, differential electrophoretic migration) is positively correlated with sequence length (i.e. number of nucleotides) (e.g. Salamon et al. 1998; Sefc et al. 2003; Stres 2006). This is consistent with the observation that the longer DNA fragments of CK7 and S72 in this study had the lowest scoring repeatability (see Chapter 2).
The size standard used in this study (LIZ1200) showed evidence of decreasing accuracy with increasing fragment length (Figure A.3), which might be expected to result in less accurate scoring of fragment lengths for longer alleles (see Pereira et al. 2001; Symonds and Lloyd 2004, among others).
In general, the reliability of post-PCR assays may also have been affected by slight differences in assay reagents and/or conditions (see Pompanon et al. 2005), as suggested 229
Appendix A.3 by the fact that the disparity between repeat assays in the present study was typically
Relative Flourescent Units
greatest when the repeats were on different plates (see Chapter 2).
Figure A.3
Peak quality of LIZ1200 size standard in relation to fragment size. Peaks
for smaller fragments are stronger, narrower and sharper than those of longer fragments.
Other Issues In the present study, there was no evidence of contamination (negative controls were clear), inconsistent poly-A addition (see Section 2.5.4), or of electrophoretic artefacts in the assays. Consequently, although each of these factors has been reported to contribute to ambiguities in the scoring of length variants in other studies (e.g. Clark 1988; Smith et al. 1995; Fernando et al. 2001; Pompanon et al. 2005), none of them appears to have been a significant issue for the intron data in the present study.
Conclusions. Despite some ambiguities with the scoring of the alleles, particularly of the larger alleles, this study has demonstrated that fragment length polymorphism at intron loci can be used to investigate population and evolutionary questions in species for which genome data are not available and funds are limited. As per other studies (e.g. Gomulski et al. 1998; Berrebi et al. 2005; Hubert et al. 2006; Atarhouch et al. 2007), these problems were addressed by binning (see Chapter 2). However, the results of this study also suggest that it would be better to focus on loci with small to medium 230
Appendix A.3 sized alleles (< 500 bp) in order to reduce the potential for scoring problems (see Chapter 2). This also has the advantage of requiring a smaller size standard, which is cheaper than the LIZ1200 (which is required for larger loci). However, such a focus will potentially make it more difficult to find highly polymorphic loci, since the level of polymorphism typically increases with locus size (see Hare 2001; Vekemans et al. 2002), which could be a major problem in species with very low levels of polymorphism (like species of Arripis).
Regardless, as with any genetic study,
genotyping errors can never be completely eliminated because neither the assays nor human factors are 100% infallible (Broquet and Petit 2003; Bonin et al. 2004; Hoffman and Amos 2005; Pompanon et al. 2005).
231
