Phenotypic characterization

3 downloads 0 Views 2MB Size Report
among strains and was recognized by winemakers before being appreciated by geneticists. In this context we constituted a S. cerevisiae wine yeast strain ...
COMPUTATIONAL APPROACHES FOR THE GENETIC AND PHENOTYPIC CHARACTERIZATION OF A SACCHAROMYCES CEREVISIAE WINE YEAST COLLECTION

Within this study, we have used a set of computational techniques to find associations between the genotypic and phenotypic characterization of autochthonous strains of Saccharomyces cerevisiae using allelic information from microsatellite loci and results from phenotypic tests. A S. cerevisiae collection, comprising 300 strains, has been obtained from winemaking environments in Portugal during the last few years. A subset of 103 strains was constituted by means of neural networks and was further characterized regarding their allelic combinations for eleven polymorphic S. cerevisiae specific microsatellites. Each strain was analysed in phenotypic screens, using taxonomic criteria (carbon and nitrogen assimilation tests, growth at different temperatures) and tests with biotechnological relevance (ethanol resistance, H2S or aromatic precursor’s formation). Globally, phenotypic results, expressed as optical density (A640) after 22 hours of growth, were in agreement with taxonomic data, although with some exceptions since few strains were capable to consume arabinose and ribose to a small extent. Taking into account that all strains derived from winemaking environments and originated from close geographic locations, the phenotypic variability was rather high. Although there were similarities between strains, each strain had a unique phenotypic profile. Based on microsatellite allelic information we trained a naïve Bayesian classifier that correctly assigned (AUC=0.81,

p0.75). These subgroups include strains with low ethanol resistance (A640

Suggest Documents