Computational Models for Prediction of Yeast Strain Potential for Winemaking from Phenotypic Profiles

et al. (2013) Computational Models for Prediction of Yeast Strain Potential for Winemaking from Phenotypic Profiles. PLoS ONE 8(7): e66523. doi:10.1371/journal.pone.0066523 Computational Models for Prediction of Yeast Strain Potential for Winemaking from Phenotypic Profiles Ine s Mendes 0 Ricardo Franco-Duarte 0 Lan Umek 0 Elza Fonseca 0 Joa o Drumonde-Neves 0 Sylvie Dequin 0 Blaz Zupan 0 Dorit Schuller 0 Joseph Schacherer, University of Strasbourg, France 0 1 CBMA (Centre of Molecular and Environmental Biology)/Department of Biology/University of Minho, Braga, Portugal, 2 Faculty of Administration, University of Ljubljana, Ljubljana, Slovenia, 3 Faculty of Computer and Information Science, University of Ljubljana, Ljubljana, Slovenia, 4 Research Center for Agricultural Technology - Department of Agricultural Sciences, University of Azores , Ponta Delgada, Sa o Miguel, Azores , Portugal , 5 INRA (Institut National de la Recherche), UMR1083, Sciences pour l'Enologie , Montpellier , France Saccharomyces cerevisiae strains from diverse natural habitats harbour a vast amount of phenotypic diversity, driven by interactions between yeast and the respective environment. In grape juice fermentations, strains are exposed to a wide array of biotic and abiotic stressors, which may lead to strain selection and generate naturally arising strain diversity. Certain phenotypes are of particular interest for the winemaking industry and could be identified by screening of large number of different strains. The objective of the present work was to use data mining approaches to identify those phenotypic tests that are most useful to predict a strain's potential for winemaking. We have constituted a S. cerevisiae collection comprising 172 strains of worldwide geographical origins or technological applications. Their phenotype was screened by considering 30 physiological traits that are important from an oenological point of view. Growth in the presence of potassium bisulphite, growth at 40uC, and resistance to ethanol were mostly contributing to strain variability, as shown by the principal component analysis. In the hierarchical clustering of phenotypic profiles the strains isolated from the same wines and vineyards were scattered throughout all clusters, whereas commercial winemaking strains tended to co-cluster. MannWhitney test revealed significant associations between phenotypic results and strain's technological application or origin. Nave Bayesian classifier identified 3 of the 30 phenotypic tests of growth in iprodion (0.05 mg/mL), cycloheximide (0.1 mg/ mL) and potassium bisulphite (150 mg/mL) that provided most information for the assignment of a strain to the group of commercial strains. The probability of a strain to be assigned to this group was 27% using the entire phenotypic profile and increased to 95%, when only results from the three tests were considered. Results show the usefulness of computational approaches to simplify strain selection procedures. - Competing Interests: The authors have declared that no competing interests exist. . These authors contributed equally to this work. Most European wine producers use commercial starter yeasts to guarantee the reproducibility and the predictability of wine quality. The advantages of fermentations containing Saccharomyces cerevisiae starter cultures relies on the fact that they are rapid and produce wine with desirable organoleptic characteristics through successive processes and harvests [1,2]. In these fermentations the winemaker has control over the microbiology of the process, because it is expected that the inoculated yeast strain predominates and suppresses the indigenous flora. Currently, there are about 200 commercial S. cerevisiae winemaking strains available, and it is a common practice among wineries to use commercial starter yeasts that were obtained in other winemaking regions. S. cerevisiae strains from diverse natural habitats harbour a vast amount of phenotypic diversity [3], driven by interactions between yeast and the respective environment. In grape juice fermentations, strains are exposed to a wide array of biotic and abiotic stressors [4], which may lead to strain selection and generate naturally arising strain diversity. Outside the wineries, this diversifying selection occurs due to unique pressures imposed after expansion into new habitats [59]. This agrees with findings showing that wine and sake strains are phenotypically more diverse than would be expected from their genetic relatedness [10]. Recent phylogenetic analyses of S. cerevisiae strains showed that the species as a whole consists of both domesticated and wild populations. DNA sequence analysis revealed that domesticated strains derived from two independent clades, corresponding to strains from winemaking and sake. Wild populations are mostly associated with oak trees, nectars or insects [1113]. Although some S. cerevisiae strains are specialized for the production of alcoholic beverages, they were derived from natural populations that were not associated with industrial fermentations. This was proposed once that the oldest lineages and the majority of variation were found in strains from sources unrelated to wine production [14]. The phenotypic diversity of S. cerevisiae strains has been explored for decades in strain selection programmes to choose the ones that enhance the wines sensorial characteristics and confer typical attributes to specific wines. These strains are used as commercial ones by winemakers to efficiently ferment grape musts and produce desirable metabolites, associated with reduced off-flavours [15,16]. Strain selection approaches are mentioned in many studies aiming to characterize S. cerevisiae isolates obtained from winemaking regions worldwide. The most relevant physiological tests refer to fermentation rate and optimum fermentation temperature, stress resistance (ethanol, osmotic and acidic), killer phenotype, sulphur dioxide (SO2) tolerance and production, hydrogen sulphide (H2S) production, glycerol and acetic acid production, synthesis of higher alcohols (e.g. isoamyl alcohol, npropanol, isobutanol), b-galactosidase and proteolytic enzyme activity, copper resistance, foam production and flocculation [17]. In our previous work [18] we evaluated the phenotypic and genetic variability of 103 S. cerevisiae strains from the Vinho Verde wine region (Northwest Portugal). We then applied several data mining procedures to estimate a strains phenotypic behaviour based on its genotypic data. We used mainly taxonomic tests and strains from winemaking environments of one geographical origin. This study was, to our best knowledge, the first attempt to computationally associate genotypic and phenotypic data of S. cerevisiae strains. We used subgroup discovery techniques to successfully identify strains with similar genetic characteristics (microsatellite alleles) that exhibited similar phenotypes. Within the present (...truncated)