Fig. 4.

H _A /H _S in outlier gene sets. In a–c, each point represents the heterozygosity at nonsynonymous sites (H _A) relative to synonymous sites (H _S) for a single gene, averaged across all samples with at least 5 heterozygous sites in the coding sequence. The size of the point is proportional to the average number of heterozygous sites relative to the length of the coding region. Best fit lines are shown for each gene set. Kernel density curves in d–f reflect the distribution of H _A /H _S values in each gene set