The number of genes showing strong signatures of positive and negative selection in conserved non-coding sites (CNCs), synonymous, and nonsynonymous sites (Nonsyn) in African Americans (AA) and European Americans (EA), and for simulated data (Sim) under the inferred AA demographic model (see Method

CNCs, synonymous, and nonsynonymous sites were analyzed both independently, and in a single concurrent run of mkprf (con), both with no fixed variance on the prior distribution of γ. Nonsynonymous sites were also analyzed by restricting to sites within human and mouse conserved regions (cons). Neutral loci simulated under the inferred AA demographic model were analyzed with a demographic correction, without a demographic correction (null demog), and assuming negative selection on synonymous sites (vs. syn+del). Total genes = genes with at least 1 fixed or polymorphic site, +/− = number of loci simulated under positive and negative selection respectively, CI = 95% credibility interval on the estimate of mean γ.