Genetic parameters and correlations between days open and ...

Livestock Science 204 (2017) 104–109

Contents lists available at ScienceDirect

Livestock Science journal homepage: www.elsevier.com/locate/livsci

Genetic parameters and correlations between days open and production traits across lactations in pasture based dairy production systems

MARK

⁎

Nicolás Frionia,b, , Gabriel Roverea,c, Ignacio Aguilard, Jorge I. Uriostea a

Facultad de Agronomía, Universidad de la República, Garzón 780, Montevideo 12900, Uruguay Animal Breeding and Genetics Group, Department of Animal Sciences, Georg-August-University, Goettingen 37075, Germany c Department of Molecular Biology and Genetics, Centre for Quantitative Genetics and Genomics, Aarhus University, Tjele, Denmark d Instituto Nacional de Investigación Agropecuaria, Ruta 48km 10 Rincón del Colorado, Canelones 90200, Uruguay b

A R T I C L E I N F O

A B S T R A C T

Keywords: Reproductive performance Dairy production Genetic correlation Heritability Holstein

The aim of this study was to estimate the additive genetic correlations, heritabilities and repeatabilities of days open (DO), milk (MY), fat (FY) and protein yields (PY), using data from pasture based dairy systems of Uruguay, and to explore whether DO should be interpreted as a different trait across lactations or as a trait with repeated measures. The database contained 500, 412 and 294 thousand records of first, second and third lactation, respectively. Cows were offspring of 7747 sires. Fertility records lower and upper limits were 42 and 250 days, respectively. In a first approach (Mdiff) we estimated variance components and covariances over lactations, assuming that the traits were different at each lactation. In a second approach (Mrep) estimations were carried out considering each trait as a repeated measure along lactations. In Mdiff, DO with a production trait was analyzed considering each lactation as a different trait. Three six-variate linear models were analyzed (DO-MY, DO-FY, DO-PY, and lactations first to third). In the Mrep procedure, DO, MY, FY and PY were analyzed together with a multiple trait repeatability model. For all models, the fixed effects were herd-year-season and lactationage classes. Animal and the permanent environment effect were included as random effects. The additive genetic correlations between DO and yield traits by Mdiff were between +0.39 and +0.78; by Mrep, they ranged from +0.44 to +0.55. Heritabilities of DO by Mdiff were between 0.04 and 0.06 and 0.05 by Mrep. The additive genetic correlations of DO between lactations ranged from +0.76 to +0.91. Heritabilities of MY, FY and PY were 0.23, 0.21 and 0.21, respectively. Repeatabilities obtained were 0.10, 0.49, 0.47 and 0.49 for DO, MY, FY and PY, respectively. We concluded that the heritability of DO was low but enough to consider the trait in selection programs. We confirmed unfavorable additive genetic correlations between DO and yield traits for the Uruguayan pasture systems, which supports the importance of considering fertility in selection programs to reduce or avoid a decline in reproduction. The additive genetic correlations of DO between lactations were high, suggesting that a plausible model should consider DO records of a given animal as repeated measures.

1. Introduction Selection in favor of production traits, e.g. milk yield (MY), has often resulted in a decline of functional traits such as the reproductive ones (Berger et al., 1981; Rauw et al., 1998; Roxström et al., 2001; Sewalem et al., 2010). Unfavorable genetic associations between production and reproductive traits come mostly from confinement production systems. Pasture-based systems, such as those predominating in Uruguay, represent a different environment, where additive genetic associations between production and reproduction might differ from those estimated in North America and Europe. For the American and Canadian Holstein populations, the reports of

⁎

additive genetic correlations between production and reproduction traits have been unfavorable (Dematawewa and Berger, 1998; Abdallah and McDaniel, 2000; Sewalem et al., 2010). Dematawewa and Berger (1998) estimated additive genetic correlations of + 0.63, + 0.58 and + 0.57 between days open (DO) - MY, DO - protein (PY) and DO - fat yield (FY), respectively, in American Holstein. In Canada, Sewalem et al. (2010) reported a genetic correlation of + 0.29 between MY and calving - first service. Abdallah and McDaniel (2000) estimated an additive genetic correlation of +0.62 between DO and fat corrected milk. Similar results have been found in Europe. For Spain, GonzálezRecio et al. (2006) reported additive genetic correlations of + 0.63, +

Correspondence to: Animal Breeding and Genetics Group, Department of Animal Sciences, Georg-August-University, Goettingen 37075, Germany. E-mail address: [email protected] (N. Frioni).

http://dx.doi.org/10.1016/j.livsci.2017.08.018 Received 14 November 2016; Received in revised form 20 June 2017; Accepted 23 August 2017 1871-1413/ © 2017 Elsevier B.V. All rights reserved.

Livestock Science 204 (2017) 104–109

N. Frioni et al.

have records in all the previous lactations. Observations of DO of 42–350 days’ range were included in the analysis. Records of DO at 250 days were considered complete, according to VanRaden et al. (2004). For production traits values within 1500–12,500 kg for MY and 50 and 350 kg for PY and FY were considered in the analysis. Additionally, observations of MY, PY or FY above or below ± 1.5 standard deviations within contemporary groups were considered outliers and deleted. Contemporary groups required a minimum of 5 observations.

0.75 and + 0.76 between DO - MY, DO - FY and DO - PY, respectively. For Norwegian red cattle, Andersen-Ranberg et al. (2005) estimated an additive genetic correlation of + 0.47 between PY and calving - first service interval. Additive genetic correlations are population parameters, affected by several factors, therefore estimates differ between countries or populations, yet unfavorable between production and reproduction. In addition, environmental and management factors have high impact on reproduction performance (Walsh et al., 2011), which explains that reproductive heritabilities are frequently below 0.10 (Dematawewa and Berger, 1998; González-Recio and Alenda, 2005; Chang et al., 2007). Uruguay represents a different scenario, where additive genetic associations between production and reproduction might differ from those estimated in North America and Europe. In Uruguay, pastures represent approximately 55% of the dry matter intake (Rovere, 2010); these systems are possibly comparable with the Irish systems, where the additive genetic correlation between calving interval and milk production along the first three lactations ranged between 0.45 and 0.66 (Olori et al., 2003). In other pasture systems, like New Zealand, the additive genetic correlation between milk and conception rate at 42 days from first to third lactation ranged from −0.013 to −0.008 (Harris et al., 2006). Pasture systems of Uruguay have been described by Lizarralde et al. (2014). Briefly, cows normally graze year-round on grass-clover pastures lasting three to four years and annual grass pastures, and varying amounts of on-farm conserved forage (maize/sorghum and grass silage) are used to complement grazing. Concentrates are strategically used during lactation to meet the nutrient requirements of their expected production level. Holstein is the predominant breed, with an average mature cow weight of 550 kg and a milk production mean of 5300 kg. Cows mostly calve in autumn and spring, to avoid calving during summer. Heifers typically calve at 27 months of age. Two different approaches have been considered for the analysis of dairy traits, (i) as a repeated measure trait or (ii) as different traits at each lactation. The second approach assumes that changes in the regulation and/or expression of the responsible genes may arise along cows’ life (Jamrozik et al., 2005; Miglior et al., 2005; González-Recio et al., 2006; Tiezzi et al., 2012). This study aims to fill a gap in the reports of additive genetic correlations between production and reproduction. The objectives were (i) to estimate the additive genetic and phenotypic correlations between DO - MY, DO - FY and DO - PY at first, second and third lactation under grazing conditions, using data from the Holstein population of Uruguay; (ii) to estimate the heritabilities and repeatabilities of each trait, and (iii) to examine whether DO should be considered as a different trait over lactations or as a repeated measure.

2.2. Statistical analysis Variables were modeled by two approaches. Firstly, each trait was considered independently at each lactation (Mdiff). Secondly, traits were considered as repeated measures across lactations (Mrep). The Mdiff model was:

Y = Xb + Za + e Where Y are the vectors of observations for linear variables (DO and MY or PY or FY) at first, second and third lactation, respectively; b is the vector of fixed effects (herd-year-season and lactation-age class); a is the vector of random animal effects; e is the random residual effect; X and Z are incidence matrices relating records to fixed and random animal effects. The Mrep model was:

Y = Xb + Za + Wp + e Where Y is the vector of observations for linear variables (DO, MY, PY and FY); b is the vector of fixed effects (herd-year-season and lactationage class); a is the vector of random animal effects; p is the vector of permanent environmental effects and non-additive genetic effects; e is the random residual effect; X, Z and W are incidence matrices relating records to fixed, animal and permanent environmental effects, respectively. For both models, it was assumed that random and residual effects are independently distributed with mean zero and variance σ2e, σ2a, respectively. In the repeatability model, it was also assumed that permanent environmental effects are independently distributed with mean zero and variance σ2p. Therefore: var(a) = Aσ2a; var(p) = Iσ2p; var(e) = Ie2e = R. Analyses were performed in a Bayesian framework, using the software Gibbs2f90 (Misztal et al., 2002). With Mdiff, a single chain of 500,000 samples after discarding the first 300,000 samples was analyzed. The sampling interval was 100, leaving 5000 samples to estimate the parameters from the posterior distributions. With Mrep, the chain analyzed was 200.000 samples long, after discarding the first 100.000 samples. The sampling interval was 10, leaving 20.000 samples to estimate the parameters from the posterior distributions. Convergence diagnostic and statistical analysis of the Markov Chain Monte Carlo sampling output were performed with CODA package (Plummer et al., 2006) of the R language/environment (R Core Team, 2014). In the Mdiff approach heritabilities (h2) were estimated as:

2. Materials and methods 2.1. Data and editing criteria Dairy production and pedigree records of 415,530 cows born between 1986 and 2009 of 1138 herds were analyzed. Production and calving information was recorded between 1990 and 2013, and 1990 and 2011, respectively. The numbers of records per lactation order are presented in Table 1. Cows were sired by 7747 bulls, with an average of 61 progenies per sire. There were 343,198 cows with both parents identified, whilst 385,429 and 399,767 animals had only sire or dam identified, respectively. The database was provided by the Instituto Nacional para el Mejoramiento y Control Lechero Uruguayo, which is the official milk recording institution of the country. The variable DO was calculated as the calving interval minus a gestation length of 280 days (Knott, 1932; VanRaden et al., 2004). MY, FY and PY are the yield of milk, fat and protein adjusted to 305 days. Lactations in the database were edited in order to discard erroneous and/or outlier records. Animals in a given lactation were required to

h2 = σa2/(σa2 + σe2) Otherwise, heritabilities (h2) were estimated as:

h2 = σa2/(σa2 + σp2 + σe2) Repeatabilities (R) were estimated as:

R = (σa2 + σp2)/(σa2 + σp2 + σe2) 105

Livestock Science 204 (2017) 104–109

N. Frioni et al.

Table 1 Number of records, statistical mean and coefficient of variation (CV) of days open (days), milk (kg), protein (kg) and fat (kg) yield at 1st, 2nd, 3rd lactation and total by Mdiff and Mrep. Days open Approach

Mdiff

Mrep

Milk yield

Fat yield

Protein yield

Lactation

n

Mean

CV

n

Mean

CV

n

Mean

CV

n

Mean

CV

First Second Third Total First Second Third Total

240,662 96,205 50,877 387,744 286,618 204,985 142,528 634,131

154 141 135 142 153 142 137 146

45.5 47.5 48.9 46.7 45.5 47.8 48.9 47.2

192,571 96,205 50,877 339,653 231,220 179,944 124,102 535,266

4897 5680 6017 5286 4877 5537 5807 5310

24.3 23.4 23.3 25.4 24.6 23.9 23.8 25.3

58,402 29,079 15,025 102,506 70,588 53,644 36,671 160,903

198.2 208.0 169.2 185.8 173.7 195.1 202.9 187.3

24.0 22.7 22.3 24.5 24.2 23.1 22.9 24.5

46,763 24,209 12,738 83,710 55,679 42,467 29,023 127,169

169.2 195.6 204.2 182.1 168.6 191.2 198.5 182.8

23.9 22.1 21.8 24.4 24.2 22.6 22.3 24.3

3. Results

Additive genetic correlations between DO and production traits ranged from + 0.44 to + 0.55 (Table 4). Phenotypic correlations between DO and production were between + 0.06 and + 0.09 (Table 4). Repeatability of DO was 0.10; the permanent effect variance was 182 days2 (Table 5). The repeatability of MY, FY and PY were 0.49, 0.47, 0.49 (Table 5). The permanent variance effect of MY, FY and PY were 1.63 × 105, 202 and 164 kg2, respectively (Table 5). The additive genetic variances were 1.44 × 105, 166 and 126 kg2 for MY, FY and PY (Table 5), respectively.

3.1. Descriptive statistics Phenotypic means of each trait along with coefficient of variation (CV) are presented in Table 1. CV of DO increased towards third lactation, whereas CV of yield traits decreased towards latter lactations. 3.2. Genetic parameters in the Mdiff approach The heritabilities of DO ranged from 0.04 to 0.06 (Table 2). Additive genetic variances of DO in first lactation ranged between 242 and 251 days2 (Table 2). In the second lactation, they were between 211 and 213 days2 and decreased to 133–153 days2 in the third lactation (Table 2). The heritabilities of MY ranged from 0.23 to 0.25 (Table 2). The heritabilities of PY and FY were of 0.28 at the first lactation, whilst in latter lactations the estimations were between 0.13 and 0.15 (Table 2). Additive genetic correlations of DO between lactations ranged from + 0.76 to + 0.91 (Table 2). For production traits, additive genetic correlations over lactations were high as well. Estimates for MY were above + 0.91, whilst genetic correlations for FY and PY were between + 0.76 and + 0.96 (Table 2). Additive genetic correlation estimates, with Mdiff, among DO and production traits over lactations varied from + 0.39 to + 0.78 (Table 3).

4. Discussion 4.1. Heritability and repeatability Heritability estimates of DO by the approaches used varied from 0.04 to 0.06 (Tables 2 and 4), being in agreement with previous results (0.04) reported by Dematawewa and Berger (1998) and Chang et al. (2007) for USA, and González-Recio and Alenda (2005) for Spain. Heritability estimates of MY in both approaches were similar 0.23–0.25) Estimates of similar magnitude were reported by Carthy et al. (2015) and González-Recio et al. (2006). Both traits, MY and DO, presented consistent heritabilities across the three lactations. Heritabilities of FY and PY in the Mdiff approach in first lactation were 0.28 and in Mrep were 0.21, which is consistent with estimates by Andersen-Ranberg et al. (2005), González-Recio et al. (2006), Abe et al. (2009) and Bastin et al. (2012). In the Mdiff approach, heritabilities of FY and PY dropped to 0.13 and 0.15 in second and third lactations. We observed that additive genetic variances moderately decreased from first to subsequent lactations, although greater changes occurred in the residual variances that

3.3. Genetic parameters in the Mrep approach Heritability estimated by the Mrep approach were 0.05, 0.23, 0.21 and 0.21 for DO, MY, FY and PY, respectively (Table 4). Additive variance of DO was 207 days2 (Table 5)

Table 2 Posterior means and standard deviations (between brackets) of additive genetic variance, heritabilities of days open (days2), milk (kg2), fat (kg2) and protein (kg2) yield, and additive genetic correlations between different lactations, obtained by Mdiff model. Additive genetic variance Traits Days open

1

Days open

2

Days open3 Milk yield Fat yield Protein yield

1 2 3

Heritabilities

Additive genetic correlations

Lact.1

Lact.2

Lact.3

Lact.1

Lact.2

Lact.3

Lact.1 − 2

Lact.1 − 3

Lact.2 − 3

242.3 (15.2) 249.5 (18.1) 251.1 (15.8) 116,588.1 (3546.9) 168.8 (8.2) 134.8 (7.2)

211.1 (18.0) 213.6 (21.0) 212.8 (19.4) 156,056.2 (5905.2) 129.7 (10.2) 91.5 (6.1)

133.4 (20.2) 152.8 (20.3) 151.2 (23.5) 190,187.1 (8854.9) 130.8 (16.5) 102.4 (10.5)

0.06 (0.004) 0.06 (0.005) 0.06 (0.003) 0.25 (0.007) 0.28 (0.014) 0.28 (0.015)

0.06 (0.005) 0.06 (0.005) 0.06 (0.005) 0.23 (0.008) 0.15 (0.011) 0.14 (0.009)

0.04 (0.005) 0.04 (0.005) 0.04 (0.006) 0.23 (0.001) 0.13 (0.015) 0.13 (0.013)

0.87 (0.030) 0.88 (0.039) 0.89 (0.031) 0.93 (0.011) 0.95 (0.009) 0.96 (0.009)

0.79 (0.055) 0.78 (0.058) 0.76 (0.055) 0.91 (0.017) 0.85 (0.041) 0.87 (0.029)

0.91 (0.043) 0.83 (0.059) 0.91 (0.041) 0.97 (0.009) 0.76 (0.072) 0.79 (0.052)

days open estimates obtained from DO-MY, DO-FY and DO-PY models, respectively.

106

Livestock Science 204 (2017) 104–109

N. Frioni et al.

Table 3 Mdiff approach posterior means of additive genetic correlations between days open-milk yield, days open-fat yield and days open-protein yield across 1st, 2nd and 3rd lactation with SE in brackets. Milk yield

Days open

Fat yield

Lactation

First

Second

Third

First

Second

Third

First

Second

Third

First

0.44 (0.033) 0.51 (0.036) 0.58 (0.061)

0.50 (0.030) 0.57 (0.039) 0.66 (0.060)

0.45 (0.044) 0.58 (0.040) 0.66 (0.067)

0.46 (0.040) 0.48 (0.050) 0.68 (0.067)

0.66 (0.037) 0.64 (0.053) 0.78 (0.058)

0.52 (0.062) 0.51 (0.067) 0.74 (0.074)

0.39 (0.043) 0.42 (0.048) 0.49 (0.078)

0.57 (0.046) 0.57 (0.047) 0.59 (0.078)

0.52 (0.073) 0.54 (0.075) 0.63 (0.069)

Second Third

misrepresented, but also there is a reduction in the number of families with records in latter lactation. This reduction in the number of families might explain the reduction in the genetic variances estimated. The increase of residual variances is also a consequence of the decline of observations at latter lactations. Individuals with no record present a zero in the dataset, meaning a missing value. The model applied predicts the missing values relying on the existing observations which represent, in relation to DO observations (240,662) at first lactation (Table 1), a 24% and 19% of observations (Table 1) at first lactation, 12% and 10% at second lactation and 5% and 6% at third lactation for FY and PY, respectively. The heritability of FY and PY obtained with the repeated measures model were both 0.21, with values close to those reported in the literature (Andersen-Ranberg et al., 2005; González-Recio et al., 2006; Abe et al., 2009). The Mrep approach also accounted for permanent effects, providing an extra source of information to predict missing values. More coherent results were obtained by Mrep, probably due to the data structure when arranging the contemporary groups, which in Mdiff resulted in a lower number of records for every single trait, particularly for FY and PY (Table 1). Although repeatability effects only explained the 9% of the variation, this estimate also accounted for permanent effects of DO variation. The model applied in the Mrep approach would be more suitable for the data structure used in this study which presented an increasing amount of missing values in latter lactations. The production traits repeatabilities were 0.49, 0.47 and 0.49 for MY, FY and PY, respectively. Comparable estimates of repeatability were reported by Carthy et al. (2015) with values of 0.56, 0.50 and 0.54 for MY, FY and PY, respectively.

Table 4 Posterior means and (standard deviations) of heritabilities (diagonal), additive genetic correlations (above diagonal) and phenotypic correlations (below diagonal) for days open, milk, fat and protein yield obtained by Mrep model.

Days open Milk yield Fat yield Protein yield

Days open

Milk yield

Fat yield

Protein yield

0.05 (0.002) 0.09 (0.003) 0.06 (0.004) 0.07 (0.004)

0.55 (0.018) 0.23 (0.004) 0.77 (0.002) 0.89 (0.001)

0.49 (0.023) 0.60 (0.011) 0.21 (0.006) 0.79 (0.002)

0.44 (0.025) 0.80 (0.006) 0.69 (0.010) 0.21 (0.006)

Table 5 Posterior means and standard deviations (between brackets) of additive genetic, permanent effect and residual variance of days open (days2), milk (kg2), fat (kg2) and protein (kg2) yield, and repeatabilities, obtained by Mrep model. Variance

Days open Milk yield Fat yield Protein yield

Protein yield

Repeatabilities

Additive genetic

Permanent effect

Residual

207 (9.0) 144,282.6 (2811.9) 166.5 (4.7) 126.3 (3.5)

181.9 (10.2) 163,542.2 (2466.2) 201.9 (4.2) 164.1 (3.0)

3558.7 (9.8) 318,701.2 (979.9) 423.4 (2.1) 299.6 (1.3)

0.10 (0.002) 0.49 (0.002) 0.47 (0.002) 0.49 (0.003)

4.2. Genetic correlation over lactations

increased 2.12 and 2.01 times at second and third lactation, respectively. The very low number of observations of fat and protein yield for second and third lactations might have resulted in difficulties to estimate missing values from prior covariances, as the groups of animals with observations for the four traits in the three lactations were scarce. Moreover, the small number of animals with observations might not be representative of the population, probably adding bias to our analysis and reflected in the values of the variances at second and third lactation of both traits. In contrast, and supporting the effect of data structure on FY and PY, the data structure of MY records allowed coherent estimates of heritability across lactations. The decrease in the additive genetic variance could be explained by two factors, the decline of observations towards later lactations and the process of animals reaching a mature state. Lower observations represent a filter consequence as only successful individuals, may present latter lactations. Farmers cull cows in early life mainly due to two reasons, reproduction and production. The latter occurs when prices are unfavorable or climate events reduce the dry matter supply, which is an important component in pasture based production systems. However, cows that are not able to be pregnant during the assigned season have higher probabilities to be culled. In this preselected sample of cows, with DO records in latter lactations, not only families are

Additive genetic correlations of DO between lactations ranged from + 0.76 to + 0.91 (Table 2). Our results were similar to those obtained by Tiezzi et al. (2012) for reproductive traits between first and second parity, such as interval from parturition to first service, interval from first service to conception and interval from parturition to conception. Additive genetic correlations reported in the literature were lower than the estimates of our study, probably because the international reports revised were based on data recorded on cows and younger animals (Muir et al., 2004; Holtsmark et al., 2008; Liu et al., 2008; Tiezzi et al., 2012). Between these two groups of animals there is a leap in body growth and energy balance (physiological factors) that leads to larger differences between young and mature life stages (Lucy et al., 1992; Loeffler et al., 1999). Given that young categories represent a large proportion of the dairy herds, the inclusion of reproductive records before first calving implies increasing the database which will improve accuracies of genetic parameters estimated. These data were not available for the present study but our results show the relevance of including reproductive records of heifers in order to improve the reproductive genetic profile. The additive genetic correlations between the production traits across lactations were high as well. The additive genetic correlation of 107

Livestock Science 204 (2017) 104–109

N. Frioni et al.

Acknowledgment

MY between lactations was the highest obtained, being above + 0.91. Correlations for FY and PY over lactations were lower than for MY, reaching values between second and third lactation of +0.76 and +0.79 for FY and PY, respectively. These lower correlations may be explained by the lower amount of observations, consequence of the data structure previously described.

This work was supported by the Agencia Nacional de Investigación e Innovación (http://www.anii.org.uy) [POS_NAC_2013_1_11539]. Special thanks to Instituto Nacional para el Mejoramiento y Control Lechero Uruguayo (http://mu.org.uy), Asociación Rural del Uruguay (http://www.aru.com.uy) and Instituto Nacional de Investigación Agropecuaria (www.inia.org.uy) for the data provided. To Agencia Nacional de Investigación e Innovación (http://www.anii.org.uy) for the financial support.

4.3. Correlation between days open and yield traits The additive genetic correlations between DO and production traits obtained with both approaches were unfavorable. Our estimation of genetic correlation between DO and MY ranged from + 0.44 to + 0.49 in Mdiff and + 0.55 in Mrep. Similar values obtained González-Recio and Alenda (2005) with a correlation of 0.63 between DO and MY for Spanish Holsteins, using a bivariate model. Analyzing other time interval reproductive traits, Sewalem et al. (2010) obtained a correlation between milk production at 90 days with calving to first service of 0.29 and with first service to conception of 0.12, in Canadian Holsteins. A preliminary study of Holstein from Uruguay, accounting up to fifth lactation with a bivariate animal model, reported an additive genetic correlation between calving interval and MY of + 0.74 (Frioni, 2012). The differences between our estimates and the obtained by Frioni (2012) might be consequence of different data editing and statistical model. The additive genetic correlations between DO and FY or PY presented a wider range in Mdiff going from + 0.46 to + 0.79 and + 0.37 to + 0.63 for FY and PY, respectively. We again suspect that those values are the consequence of poor data structure and bias. The additive genetic correlation between DO and FY or PY estimated by Mrep presented a small range, going from + 0.49 to + 0.44, respectively. González-Recio et al. (2006) reported values of 0.75 and 0.76 between DO-FY and DO-PY, respectively, using bivariate models. Andersen-Ranberg et al. (2005) with Norwegian dairy cattle estimated an additive genetic correlation between calving to first service and PY of 0.47 with multivariate linear sire models.

References Abdallah, J.M., McDaniel, B.T., 2000. Genetic parameters and trends of milk, fat, days open, and body weight after calving in North Carolina experimental herds. J. Dairy Sci. 83, 1364–1370. http://dx.doi.org/10.3168/jds.S0022-0302(00)75004-1. Abe, H., Masuda, Y., Suzuki, M., 2009. Relationships between reproductive traits of heifers and cows and yield traits for Holsteins in Japan. J. Dairy Sci. 92, 4055–4062. http://dx.doi.org/10.3168/jds.2008-1896. Andersen-Ranberg, I.M., Klemetsdal, G., Heringstad, B., Steine, T., 2005. Heritabilities, genetic correlations, and genetic change for female fertility and protein yield in Norwegian Dairy Cattle. J. Dairy Sci. 88, 348–355. http://dx.doi.org/10.3168/jds. S0022-0302(05)72694-1. Bastin, C., Berry, D.P., Soyeurt, H., Gengler, N., 2012. Genetic correlations of days open with production traits and contents in milk of major fatty acids predicted by midinfrared spectrometry. J. Dairy Sci. 95, 6113–6121. http://dx.doi.org/10.3168/jds. 2012-5361. Berger, P.J., Shanks, R.D., Freeman, A.E., Laben, R.C., 1981. Genetic aspects of milk yield and reproductive performance. J. Dairy Sci. 64, 114–122. http://dx.doi.org/10. 3168/jds.S0022-0302(81)82535-0. Carthy, T.R., Ryan, D.P., Fitzgerald, A.M., Evans, R.D., Berry, D.P., 2015. Genetic relationships between detailed reproductive traits and performance traits in HolsteinFriesian dairy cattle. J. Dairy Sci. 99, 1–12. http://dx.doi.org/10.3168/jds.20159825. Chang, Y.M., González-Recio, O., Weigel, K.A., Fricke, P.M., 2007. Genetic analysis of the twenty-one-day pregnancy rate in US Holsteins using an ordinal censored threshold model with unknown voluntary waiting period. J. Dairy Sci. 90, 1987–1997. http:// dx.doi.org/10.3168/jds.2006-359. Dematawewa, C.M., Berger, P.J., 1998. Genetic and phenotypic parameters for 305-day yield, fertility, and survival in Holsteins. J. Dairy Sci. 81, 2700–2709. http://dx.doi. org/10.3168/jds.S0022-0302(98)75827-8. Frioni, N., 2012. Estimación de la Heredabilidad del Inervalo Entre Partos y su Correlación Genética con Producción de Leche en Ganado Holando Uruguayo. Universidad de la República. González-Recio, O., Alenda, R., 2005. Genetic parameters for female fertility traits and a fertility index in Spanish dairy cattle. J. Dairy Sci. 88, 3282–3289. http://dx.doi.org/ 10.3168/jds.S0022-0302(05)73011-3. González-Recio, O., Alenda, R., Chang, Y.M., Weigel, K.A., Gianola, D., 2006. Selection for female fertility using censored fertility traits and investigation of the relationship with milk production. J. Dairy Sci. 89, 4438–4444. http://dx.doi.org/10.3168/jds. S0022-0302(06)72492-4. Harris, B., Pryce, J.E., Xu, Z.Z., Montgomery, G.M., 2006. Development of new fertility breeding values in the dairy industry. Proc. New Zeal. Soc. Anim. Prod., 66, 6. Holtsmark, M., Heringstad, B., Madsen, P., Ødegård, J., 2008. Genetic relationship between culling, milk production, fertility, and health traits in Norwegian red cows. J. Dairy Sci. 91, 4006–4012. http://dx.doi.org/10.3168/jds.2007-0816. Jamrozik, J., Fatehi, J., Kistemaker, G.J., Schaeffer, L.R., 2005. Estimates of genetic parameters for Canadian Holstein female reproduction traits. J. Dairy Sci. 88, 2199–2208. http://dx.doi.org/10.3168/jds.S0022-0302(05)72895-2. Knott, J.C., 1932. A study of the gestation period of Holstein-Friesian cows. J. Dairy Sci. 15, 87–98. http://dx.doi.org/10.3168/jds.S0022-0302(32)93391-8. Liu, Z., Jaitner, J., Reinhardt, F., Pasman, E., Rensing, S., Reents, R., 2008. Genetic evaluation of fertility traits of dairy cattle using a multiple-trait animal model. J. Dairy Sci. 91, 4333–4343. http://dx.doi.org/10.3168/jds.2008-1029. Lizarralde, C., Picasso, V., Rotz, C.A., Cadenazzi, M., Astigarraga, L., 2014. Practices to Reduce Milk Carbon Footprint on Grazing Dairy Farms in Southern Uruguay: Case Studies. Sustain. Agric. Res. 3. http://dx.doi.org/10.5539/sar.v3n2p1. Loeffler, S.H., de Vries, M.J., Schukken, Y.H., de Zeeuw, A.C., Dijkhuizen, A.A., de Graaf, F.M., Brand, A., 1999. Use of AI technician scores for body condition, uterine tone and uterine discharge in a model with disease and milk production parameters to predict pregnancy risk at first AI in Holstein dairy cows. Theriogenology 51, 1267–1284. http://dx.doi.org/10.1016/S0093-691X(99)00071-0. Lucy, M.C., Staples, C.R., Thatcher, W.W., Erickson, P.S., Cleale, R.M., Firkins, J.L., Clark, J.H., Murphy, M.R., Brodie, B.O., 1992. Influence of diet composition, dry-matter intake, milk production and energy balance on time of post-partum ovulation and fertility in dairy cows. Anim. Prod. 54, 323–331. http://dx.doi.org/10.1017/ S0003356100020778. Miglior, F., Muir, B.L., Van Doormaal, B.J., 2005. Selection indices in Holstein cattle of various countries. J. Dairy Sci. 88, 1255–1263. http://dx.doi.org/10.3168/jds. S0022-0302(05)72792-2. Misztal, I., Tsuruta, S., Strabel, T., Auvray, B., Druet, T., Lee, D.H., 2002. BLUPF90 and related programs (BGF90), In: Proceedings of the 7th World Congress on Genetics

5. Conclusions Our study detected an important phenotypic variation in DO under grazing conditions. Heritability estimates of DO were low and did not vary between the Mdiff and Mrep approaches 0.03–0.06 by Mdiff; 0.05 by Mrep). The repeatability of DO indicated that permanent effects represent a source of variation for the trait. Considering in addition that the additive genetic correlations of DO across lactations were favorable and high, we consider that DO should be analyzed as a single trait with repeated measures. The genetic correlations among DO and yield traits over lactations were between + 0.39 and + 0.78 with the Mdiff approach; with the repeatability model, the same parameter ranged from + 0.44 to + 0.55. We conclude that estimating the additive genetic correlations through the repeatability model was a more reliable approach, considering the permanent effect accounted in this model, which was proved to be a source of variation. The additive genetic correlation between DO and production traits estimated in the present study from pasture systems did not differ from the estimates reported for in-door production systems. This indicates that there is a significant genetic antagonism between production and reproduction in pasture-based dairy system, which reinforces the relevance of including reproduction traits in dairy breeding programs.

Conflict of interest There is no conflict of interest for this manuscript. 108

Livestock Science 204 (2017) 104–109

N. Frioni et al.

Roxström, A., Strandberg, E., Berglund, B., Emanuelson, U., Philipsson, J., 2001. Genetic and environmental correlations among female fertility traits and milk production in different parities of Swedish red and white dairy cattle. Acta Agric. Scand. Sect. A Anim. Sci. 51, 7–14. http://dx.doi.org/10.1080/090647001300004745. Sewalem, A., Kistemaker, G.J., Miglior, F., 2010. Relationship between female fertility and production traits in Canadian Holsteins. J. Dairy Sci. 93, 4427–4434. http://dx. doi.org/10.3168/jds.2009-2915. Tiezzi, F., Maltecca, C., Cecchinato, A., Penasa, M., Bittante, G., 2012. Genetic parameters for fertility of dairy heifers and cows at different parities and relationships with production traits in first lactation. J. Dairy Sci. 95, 7355–7362. http://dx.doi.org/10. 3168/jds.2012-5775. VanRaden, P.M., Sanders, A.H., Tooker, M.E., Miller, R.H., Norman, H.D., Kuhn, M.T., Wiggans, G.R., 2004. Development of a national genetic evaluation for cow fertility. J. Dairy Sci. 87, 2285–2292. http://dx.doi.org/10.3168/jds.S0022-0302(04) 70049-1. Walsh, S.W., Williams, E.J., Evans, A.C.O., 2011. A review of the causes of poor fertility in high milk producing dairy cows. Anim. Reprod. Sci 123. pp. 127–138. http://dx.doi. org/10.1016/j.anireprosci.2010.12.001.

Applied to Livestock Production, pp. 21–22. Muir, B.L., Fatehi, J., Schaeffer, L.R., 2004. Genetic relationships between persistency and reproductive performance in first-lactation Canadian holsteins. J. Dairy Sci. 87, 3029–3037. http://dx.doi.org/10.3168/jds.S0022-0302(04)73435-9. Olori, V.E., Pool, M.H., Calus, M.P.L., Cromie, A.R., Veerkamp, R.F., 2003. Joint evaluation of survival and fertility in dairy cattle with a linear model. Interbull Bull. 30, 20–24. Plummer, M., Best, N., Cowles, K., Vines, K., 2006. CODA: convergence diagnosis and output analysis for MCMC. R News. 〈https://cran.r-project.org/web/packages/ coda/coda.pdf〉. R Core Team, 2014. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. R Found. Stat. Comput. Rauw, W., Kanis, E., Noordhuizen-Stassen, E., Grommers, F., 1998. Undesirable side effects of selection for high production efficiency in farm animals: a review. Livest. Prod. Sci. 56, 15–33. http://dx.doi.org/10.1016/S0301-6226(98)00147-X. Rovere, G., 2010. Derivación de valor económico para características de producción y fertilidad en sistemas lecheros de base pastoril. Universidad de la República Facultad de Agronomía.

109