Understanding the rules laying behind the natural phenotypic variation has been a key point of modern genetics for decades. However, it is still difficult to precisely address and dissect the molecular bases underlying complex traits. Today, a better understanding of the genetic architecture of traits requires a precise estimation of the genetic components governing phenotypes at a species-wide level.

In this context, we took advantage of the large set of 1,011 natural Saccharomyces cerevisiae isolates that we recently sequenced. We selected a set of 55 isolates as genetically diverse as possible to generate a diallel cross panel of 3,025 hybrids. These hybrids were then phenotyped on 49 stress related traits resulting in 148,225 cross/trait combinations. The results clearly showed that although phenotypic variance is mostly governed by additivity, 30% of this variance can be explained by non-additive phenomena. This is confirmed by the fact that a majority of complete dominance is observed in 25% of the traits. The dataset we generated also allowed us to perform genome-wide association studies (GWAS) to uncover variants responsible for the tested phenotypes. Interestingly, 2,156 significantly associated variants were found and among them 12% are present in less than 5% of the 1,011 population.

It clearly shows that those so-called rare variants represent an important source of phenotypic variance and can be mapped using GWAS on a diallel panel. To complete this view, we are currently looking at the phenotypic distribution and segregation in the progeny to uncover the phenotypic expressivity variation across genetic backgrounds.

Visit the NEUROGENYCS page