The effects of selfing on multi-step adaptation

Macroevolutionary studies have estimated higher extinction rates of self-compatible lineages than self-incompatible ones. A leading explanation is that selfing may prevent adaptation, since models show that selfing can inhibit the fixation of adaptive alleles at a single locus (1-step adaptation). However, adaptation often involves changes at multiple loci (multi-step adaption), but the effects of selfing remain unclear because selfing increases homozygosity, which affects selection intensity, the effective population size, and the effective recombination rate. By modeling using population genetic models, I investigate the effects of selfing on adaption requiring fixation of 2 adaptive alleles, I show that intermediate selfing rates generally promote adaption, by increasing the fixation probability of the double-mutant haplotype once it is generated. In constant-sized populations, selfing increases the rate of adaptation through the fixation of new mutations even when both alleles are dominant. In demographically declining populations, the rescue probability rises sharply as the selfing rate increases from zero, but quickly drops to be low when it approaches 1.0. These findings are at odds with the hypothesis that higher extinction rates of self-compatible lineages result from reduced adaptive potential but may help explain why some studies have failed to detect relaxation of selection in selfers and also the prevalence of mixed-mating systems.