Dr. Gilbert obtained her PhD from the University of British Columbia in 2016, studying theoretical population genetics and the impact of demography on evolutionary processes and inferences. Her research broadly includes both theoretical and empirical data analysis in topics of evolutionary biology, including population structure, effective population size, local adaptation, and mutation load. She is currently a postdoctoral fellow at the University of Lausanne, Switzerland. More information is available on her website: http://kjgilbert.github.io/
Abstract
Linked selection is a major driver of genetic diversity. Selection against deleterious mutations removes linked neutral diversity (background selection [BGS]), creating a positive correlation between recombination rates and genetic diversity. Purifying selection against recessive variants, however, can also lead to associative overdominance (AOD), due to an apparent heterozygote advantage at linked neutral loci that opposes the loss of neutral diversity by BGS. Zhao and Charlesworth (2016) identified the conditions under which AOD should dominate over BGS in a single-locus model and suggested that the effect of AOD could become stronger if multiple linked deleterious variants co-segregate. We present a model describing how and under which conditions multi-locus dynamics can amplify the effects of AOD. We derive the conditions for a transition from BGS to AOD due to pseudo-overdominance, i.e., a form of balancing selection that maintains complementary deleterious haplotypes that mask the effect of recessive deleterious mutations. Simulations confirm these findings and show that multi-locus AOD can increase diversity in low-recombination regions much more strongly than previously appreciated. While BGS is known to drive genome-wide diversity in humans, the observation of a resurgence of genetic diversity in regions of very low recombination is indicative of AOD. We identify 22 such regions in the human genome consistent with multi-locus AOD. Our results demonstrate that AOD may play an important role in the evolution of low-recombination regions of many species.
