Fig. 1
Simulation of recurrent positive selection (model A) under various population sizes and selection coefficients. A sequence of length L = 1,000 contains 230 synonymous sites over which the level of polymorphism is calculated, which models the HA1 segment of H3N2 virus. As we reduced L b (the number of beneficial nonsynonymous sites) with increasing s and N, the numbers of adaptive substitutions per year are similar (k ≈ 1.3) in all cases. a. Per-segment synonymous diversity (πs) as a function of population size with different assumptions on selection coefficient s for beneficial mutations. πs obtained from HA1 is indicated by a dashed line. b. The rate of soft sweeps (f soft) in the same simulation replicates