The composition of the find more African American, U.S. Caucasian and U.S. Hispanic populations, and the extent of the diversity within each of the ancestry groups that contribute to them, are reflected in pairwise comparisons performed for (a) each population sample and (b) all samples ascribed to each of the four biogeographic ancestry categories.

Fig. 2 displays histograms of pairwise comparisons for both the full mtGenome and the CR only, for each of the three populations and three of the four ancestry groups, plotted by the proportion of comparisons performed to normalize for the differing sample sizes. The average number of pairwise differences for each of these sets of comparisons are reported in Table S6. When the entire mtGenome was considered, the U.S. Caucasian population sample (Fig. 2b) and the haplotypes of West Eurasian ancestry (Fig. 2e) had asymmetrical bimodal pairwise distributions, with the first, smaller peak representing the comparisons between recently diverged lineages in the dataset, and the second, larger peak representing the comparisons between more distantly related haplotypes. When these same analyses were performed with the comparison restricted to the CR (Fig. 2h and k), the distributions were unimodal and Poisson-like (though still significantly different from a Poisson distribution; p < 0.0001 Vemurafenib supplier for both). For the U.S. Hispanic dataset,

Fig. 2c displays an asymmetrical bimodal distribution similar to the U.S. Caucasians, but with a substantial tail to the right that represents comparisons to and between the African ancestry haplotypes present in the population sample. The Native American ancestry comparisons ( Fig. 2f and l) are sharply bimodal and more symmetrical, reflecting the origins of Native Americans and the genetic distance between aminophylline the haplotypes in this sample set (primarily, haplogroups A and B from macrohaplogroup N, and haplogroups C and D from

macrohaplogroup M). The comparisons between these haplotypes based on the CR alone ( Fig. 2l) are the only CR pairwise distribution that closely mirrors the shape of the distribution based on the full mtGenome. In contrast to the other sample sets, comparisons of both the African American population sample and the African ancestry lineages for the complete mtGenome resulted in multimodal distributions ( Fig. 2a and d) and high average pairwise numbers of differences (Table S6). In comparison to the U.S. Caucasian and U.S. Hispanic populations, fewer of the African American haplotypes are highly similar to one another across the entire mtGenome, and a much greater number are genetically very distant. Consistent with results from previous studies of African American population samples [7], [46], [48], [49] and [50], the distributions for these two comparisons underscore the extensive mtDNA diversity that exists within the African ancestry component of U.S. populations.