Soft Sweeps III: The Signature of Positive Selection from Recurrent Mutation

Abstract

Polymorphism data can be used to identify loci at which a beneficial allele has recently gone to fixation, given that an accurate description of the signature of selection is available. In the classical model that is used, a favored allele derives from a single mutational origin. This ignores the fact that beneficial alleles can enter a population recurrently by mutation during the selective phase. In this study, we present a combination of analytical and simulation results to demonstrate the effect of adaptation from recurrent mutation on summary statistics for polymorphism data from a linked neutral locus. We also analyze the power of standard neutrality tests based on the frequency spectrum or on linkage disequilibrium (LD) under this scenario. For recurrent beneficial mutation at biologically realistic rates, we find substantial deviations from the classical pattern of a selective sweep from a single new mutation. Deviations from neutrality in the level of polymorphism and in the frequency spectrum are much less pronounced than in the classical sweep pattern. In contrast, for levels of LD, the signature is even stronger if recurrent beneficial mutation plays a role. We suggest a variant of existing LD tests that increases their power to detect this signature.

Synopsis

Populations adapt to their environment through fixation of beneficial alleles. Such fixation events leave a signature in neutral DNA variation of the population. An accurate description of this signature, also called a selective sweep, can be used to identify genes that have been involved in recent adaptations. The classical model of a selective sweep assumes that the beneficial allele was created only once by mutation, whereas the authors have shown, in a previous paper, that this assumption does not always hold. If a substitution involves multiple copies of an allele that have originated by independent mutation, it leads to a different signature, which the authors call a soft selective sweep. In this study, Pennings and Hermisson use analytical tools and coalescent simulations to describe this soft-sweep pattern. They show that this pattern is characterized by strong linkage disequilibrium. They also analyze the power of standard tests of neutrality to detect this pattern and suggest a variant of existing linkage-disequilibrium–based tests that increase the power to detect positive selection in the form of a soft selective sweep.