Impacts of pleiotropy and migration on repeated genetic adaptation

Abstract Observations of genetically repeated evolution (repeatability) in complex organisms are incongruent with the Fisher-Orr model, which implies that repeated use of the same gene should be rare when mutations are pleiotropic (i.e., affect multiple traits). When spatially divergent selection occurs in the presence of migration, mutations of large effect are more strongly favoured, and hence repeatability is more likely, but it is unclear whether this observation is limited by pleiotropy. Here, we explore this question using individual-based simulations of a two-patch model incorporating multiple quantitative traits governed by mutations with pleiotropic effects. We explore the relationship between fitness trade-offs and repeatability by varying the alignment between mutation effect and spatial variation in trait optima. While repeatability decreases with increasing trait dimensionality, trade-offs in mutation effects on traits do not strongly limit the contribution of a locus of large effect to repeated adaptation, particularly under increased migration. These results suggest that repeatability will be more pronounced for local rather than global adaptation. Whereas pleiotropy limits repeatability in a single-population model, when there is local adaptation with gene flow, repeatability can occur if some loci are able to produce alleles of large effect, even when there are pleiotropic trade-offs. Article summary Classical evolutionary theory predicts that genetically repeated evolution should be vanishingly rare in organisms where mutations affect multiple traits. In this article, we use simulations to study such pleiotropic mutations, and explore their effects on local adaptation in two patches under divergent selection. We find that migration between patches increases the likelihood of repeated evolution, even when there are fitness trade-offs imposed by pleiotropy.