We evaluated whether there is an excess of common SNPs with large allele frequency differences between Icelandic subpopulations, using data from 14,313 individuals with majority ancestry from one of 11 Icelandic regions (Table 1). For each Icelandic region, we computed the distribution of allele frequency differences between that region and the union of all other regions, expressed as a χ2 (1 d.o.f.) statistic under a model of neutral genetic drift. This computation accounts for related individuals (see Materials and Methods). P-P plots for each region r () are displayed in Figure 4. For each region, there was no excess of markers with large frequency differences versus other regions. Averaging across computations for each of 11 regions, 0.008% of markers had a P-value less than 0.0001, roughly matching the expected distribution. The most significant P-value was 3×10−6, a value that is not statistically significant after correcting for the number of SNPs and regions tested. These results are consistent with the hypothesis that the divergence time of Icelandic regions has been too short for differential selective forces to have had a significant impact on allele frequencies.
