Various observations argue for a role of adaptation in recent human evolution, including results from genome-wide studies and analyses of selection signals at candidate genes. . . On a broad scale, the geographic distribution of putatively selected alleles almost invariably conforms to population clusters identified using randomly chosen genetic markers. Given this structure, there are surprisingly few fixed or nearly fixed differences between human populations. Among the nearly fixed differences that do exist, nearly all are due to fixation events that occurred outside of Africa, and most appear in East Asia (my emphasis).The authors focus on distributions associated with three genes, the "KIT ligand gene," and the "SLC24A5 gene," both apparently associated with light skin pigmentation, and a "nonsynonymous SNP in the MC1R gene." The distribution patterns for all three are presented in Figure 4:
Figure 4. Global allele frequencies and haplotype patterns at three genes with signals of positive selection. The left-hand column shows pie charts of allele frequencies (blue ancestral, red derived) across the HGDP populations for: (A) a SNP upstream of KITLG (rs1881227); and for nonsynonymous SNPs in (B) SLC24A5 (rs1426654; data from ), and (C) MC1R (rs885479). The right-hand column shows a representation of haplotype patterns for 500 kb around each gene, in each case centered on the SNP displayed in the pie charts. Each box represents a single population, and observed haplotypes are plotted as thin horizontal lines, using the same haplotype coloring for all populations (see Methods and ). In all three cases the derived allele plotted in the pie charts is found mainly on the red haplotype.
The gap we've been discussing can clearly be seen in map B, representing the SLC24A5 gene (possibly associated with skin color). Significantly, blue, found predominantly in Africa, East Asia and Melanesia, represents the ancestral form of the allele, and red, found exclusively in Europe, the Middle East and South Asia, represents the more recent, derived, allele. The gap can also be seen in the rectangular graphs on the right, with the patterns for Europe, the Middle East and S. Asia clearly contrastive with those for Africa, E. Asia, Oceania and the Americas.
Another paper, also dealing with worldwide distributions of specific genes, has already been discussed on this blog (see Post 265), as an example of how not to interpret simple correlations: Linguistic tone is related to the population frequency of the adaptive haplogroups of two brain size genes, ASPM and Microcephalin, by Dan Dediu and D. Robert Ladd. The authors found a correlation between the distribution of two genes thought to be associated with brain size, and the worldwide distribution of tone languages. The most convincing correlation is between the distribution of ASPM-D and tone language, as seen in the following maps:
Though the pattern shown on the ASPM-D map isn't as clear as the one on the tone language map, the authors claim to have found a statistically significant correlation between the two -- and on that basis, concluded that there must be a cause and effect relation between the presence of the derived forms of the genes and the absence of tone language. As I argued in Post 265, their cause and effect inference is highly questionable, since they failed to consider the fact that the gene distributions themselves must have a cause. And in the light of all the evidence we've been considering here, it looks very much like the correlation found by Dediu and Ladd stems from the likelihood that both distributions have a common cause, the very cause their mathematical model ruled out as unlikely: demic diffusion over a continuous geographical area, from Africa all the way to East Asia -- a pattern masked by the same gap we've been finding, over and over, as we consider evidence, both cultural and genetic, associated with the Out of Africa migration.