Sometimes, when human clinical studies are too difficult to undertake, and animal studies (such as the intact mouse) are similarly difficult to perform and tease out important data from all “background noise,” it is useful to study yeast cells in culture. They have a sufficiently complex genome, combined with many traits that can be quantified. The attached study of genetic variation in yeast has identified crucial
quantitative trait loci (QTLs) that suppress specific effects of variability at multiple other loci (DNA sites). These loci are proven to be essential to accurately modeling yeast growth in response to different environments.
When the combined phenotypic effect of alleles at two or more loci deviates from the sum of their individual effects, this is referred to as “gene x gene interaction,” or epistasis. Most biological traits are regulated by a complex interplay between multiple genes and environmental factors. Despite this underlying complexity –– data and theory have shown we expect that most of the genetic variance in a population will be additive (i.e. equal contribution from the gene’s alleles on both chromosomes).
The apparent contradiction between the complexity of biological mechanisms that determine quantitative traits and observation that most genetic variance can be captured by an additive model –– has led to a long-standing debate in genetics: does the predominant role of additive genetic variance mean that strictly additive models are always sufficient to describe the relationship between the genotype and phenotype of an organism or could there be added value in explicitly modeling genetic interactions, despite the lower levels of epistatic genetic variance?
Quantitative genetics theory predicts –– and many lines of data support –– the notion that most genetic variance in populations is additive in its inheritance. Herein the authors describe networks of capacitating genetic interactions that contribute to QTL variation in a large yeast intercross population. The additive variance explained by individual loci in a network is highly dependent on allele frequencies of the interacting loci. Modeling of phenotypes for multi-locus genotype classes in the epistatic networks is often improved by accounting for such interactions. Authors discuss the implications of these results with regard to attempts to dissect genetic architectures and to predict individual phenotypes and long-term responses to selection under varying environmental conditions.
Nature Genet April 2o17; 49: 497–503 [article] and 486–488 [ed.]