One of the most interesting results in the preprint on ancient European genetics (or more accurately, the ethnogenesis of Europeans in a genetic sense) is the fact that the ~8,000 year old hunter-gatherer sample from Luxembourg had a GG genotype on the SLC24A5 locus. Actually, interesting isn’t the right word, shock, and frankly a little skepticism is more precise. The reason for my reflexive incredulity is that the GG genotype is very much the minor variant in Western Eurasia, and extremely rare among unadmixed Europeans. Europeans have such a high fraction of the A allele that some population genetic statistics to test for selection at a locus are not viable, because there’s not enough variation segregating in that region. This allele also is present outside of Europe, with the A allele being the major variant in South Asians, albeit at a lower fraction, verging on ~50% or less in some South Indian groups. Yet it is not entirely implausible that this allele only swept to fixation over the past 8,000 years in Europe looking at the genomic features* of the region in which it is embedded.
I want to make more concrete why this result is a pretty big deal. If you look at the 1000 Genomes data you have results for British, Finnish, Tuscan, and Spanish individuals, as well as a well characterized sample of white Utahans of Northwest European heritage. There is also a less well characterized pooled data set of “European Americans.” Here are the genotype counts by population:
Yesterday on Twitter I suggested that I’d want at least 10,000 individuals of unadmixed Northern European ancestry before I might take a bet that I’d find someone with a GG genotype. I don’t think I was exaggerating. The sample size might be one, but the fact that the individual was homozygous for GG implies to me that the G allele was present at a far higher fraction in Northern Europe 8,000 years ago than today. In contrast the LBK farmer individual was AA on SLC24A5. Why this matters functionally is that no matter how you look at it, when comparing Europeans and dark skinned populations (e.g., Africans, South Indians, and Australasians) this locus is the one that explains the highest proportion of the variation on pigmentation of any gene. Comparing simply people of African ancestry and Europeans the variation at this gene accounts for on the order of ~1/3 of the difference.** I myself have the “European” AA genotype, with most of my other large effect loci being of the “dark” correlated alleles. The pigmentation difference between a Sub-Saharan African and myself is probably accounted for just by this locus alone. But a twist on this story is that the hunter-gatherer also exhibited the genotype associated with blue eyes in Europeans. In contrast, the farmer genotype was the one not correlated with blue eyes. On another locus which is not quite fixed for a derived light encoding variant, but very close in Europeans (and found in much lower proportions in other West Eurasians), SLC45A2, it looks as if both the hunter-gatherer and the farmer carry the modal European form.
Rather that squeezing too much more out of a few samples, I want to posit that these results increase the plausibility that the suite of genetic variants across many loci which are often diagnostic of the complexion of Northern Europeans are a function of a combination of admixture and then selection within the resultant Northwest European lineages. It seems plausible that independent selection events were occurring across these groups, and with admixture more novel variants were present in the combined population which allowed for a skew even further along the phenotypic continuum, toward the physiological limit (at least for non-albinos). Though it looks like the majority of the ancestry of Northern Europeans, especially populations around the coastal East Baltic region, derive from hunter-gatherer groups indigenous to the continent (i.e., pre-Holocene), if they were not fixed for the derived variant on SLC24A5 it seems implausible that these ur-Europeans were defined by the rosy complexions which are archetypical for Northern Europeans . This is part of the broader picture whereby the phenotypically salient population clusters we see around us today, as if they are Platonic ideals of underlying racial forms, may themselves be phenomena distinctive to the Holocene .
* A large correlated block of markers which seem to have risen in frequency recently and rapidly within the population.
** Northeast Asians have their own distinct mutations which confer light skin.