We examined the subset of SNPs (862) from the ENCODE sequence data that were present at low frequency (two to six occurrences of the alleles) and were also observed in more than one population. These are of special interest as they would be most likely to include examples of independent mutations that occurred since the populations diverged, as opposed to each observed allele being descended from a single ancestral event. In the majority of cases (93%) the rare variants at each site occurred on the same haplotype background, consistent with a single origin, and their current distribution reflects drift. The remaining 51 sites (7%) were observed to have alleles that occurred in more than one haplotype. Furthermore, the different haplotypes occurred in different populations for all except one site. These 51 sites are therefore candidates for independent occurrence of mutation at the same site (Supplementary Information and Supplementary Table 7).
