FIGURE 2. Demography and selection can cause a proportional excess of non-synonymous SNPs in Europeans.

a, b, Results of forward simulations of a population that expanded (AA 2 in Supplementary Table 2), to represent the AA population and a population that experienced a bottleneck to represent the EA population (EA 1 in Supplementary Table 2). a, Distribution of the proportion of non-synonymous SNPs segregating in samples simulated under European (dashed curve) and African (solid curve) demographic models. Vertical lines show the observed proportions in the Applera data set. b, Distribution of selection coefficients for simulated SNPs in the AA (white bars) and the EA (grey bars) samples. The labels on the x axis are the more negative limits of the bins. Error bars denote 95% intervals on the proportion of SNPs in each group. c–e, Expected distribution of SNPs over time during a population expansion (AA 2, solid lines), a long, mild bottleneck (EA 1, dashed lines) and a short, severe bottleneck (EA 6, dotted lines). Time moves forward from left to right. Solid vertical lines indicate when the populations changed size. Further details are given in Supplementary Table 2. c, The number of non-synonymous SNPs. d, The number of synonymous SNPs. e, The proportion of non-synonymous SNPs.