The efficacy of QTL analysis to predict adaptive variation: a test using experimentally evolved populations of yeast
Authors: Helen Murphy; Benjamin Epley; Brianna Meeks; Juliana Salcedo
Affiliation: William and Mary
Keywords: Experimental evolution
A major goal of evolutionary biology is to identify adaptive genetic variation and understand its fate in a population. Quantitative trait locus (QTL) analysis is a common approach used to uncover variants associated with adaptive traits. It is unclear whether QTL hypothesized to underlie traits of interest are in fact those that will be favored by natural selection, as complex life histories, epistasis, and pleiotropy may affect the strength and direction of selection on these loci. We tested whether variants implicated in a QTL analysis increased in frequency when under selection using the yeast Saccharomyces cerevisiae as a model. The ability to adhere to plastic, which is the first step in biofilm formation and a clinically-relevant trait, was investigated in a highly-heterozygous strain isolated from a medical setting (YJM311). First, a bulk segregant analysis (BSA) was used to identify plastic adherence QTL; this assay implicated variants in a number of genes known to be associated with other biofilm-related traits, including FLO8 and FLO11. Next, experimental evolution was performed on replicate sexual and asexual populations for 500 generations while selecting for the ability to adhere to plastic. In order to identify loci under selection, populations were subject to whole-population, whole-genome sequencing at 8 time points throughout the experiment. Thousands of variants were found to be under selection, many likely related to adapting to the experimental conditions and not just increasing the ability to adhere to plastic. When the two datasets were compared, nearly all the loci with the strongest signal in the BSA increased in frequency throughout the selection experiment, suggesting the QTL analysis successfully uncovered evolutionarily-relevant loci. However, most of these variants did not fix in the experimental populations. These results suggest that the loci uncovered in the QTL analysis were subject to a more complex process than a simple selective sweep of adaptive loci and that populations with abundant segregating variation may not conform to simple evolutionary models.