Selection for rs3827760 at EDAR (“shovel-shaped” incisor SNP) during Holocene around the “Ring of Fire”

If you have been reading my blog will you be familiar with the SNP rs3827760
within the EDAR gene. This mutation has high derived frequencies in East Asians and is associated with a suite of physical characteristics. Most famously, the thickness of hair shaft and “shovel-shaped” incisors (a phenotype also found in Neanderthals). So the reason people of East Asian ancestry seem to have very thick straight hair is that their hair strands are actually thicker due to the new variant.

Almost all Africans, West Eurasians, and South Eurasians lack the derived variant. Those populations outside of East Asia which have it in appreciable frequencies, whether it be Munda tribal people in India or Finns in Northern Europe, always have relatively recent East Asian ancestry. The fraction of the derived allele is usually easily inferred from genome-wide East Asian ancestry and source population fraction (southern East Asians have a lower fraction than northern ones).

But there’s another modern group* of people with high frequencies of the derived variant: people of Amerindian heritage. This is reasonable because East Asians and Amerindians share common ancestry, at least in part, going back ~25,000 years ago. The ALFRED database actually has the largest coverage of the New World for this marker that I know of. One inference you can make is that many Amerindian groups were fixed or nearly fixed for the derived variant before some European admixture. For example, the Maya carry ~5% of the ancestral variant, but those samples are known to have a small but significant amount of European admixture (curiously, the derived variant hasn’t swept to fixation in many populations; that implies to me that the phenotypic target of selection has a dominant genetic expression).

So this section of a new ancient DNA paper, Reconstructing the Deep Population History of Central and South America, jumped out at me:

Our data show that a variant in EDAR that affects tooth shape, hair follicles and thickness, sweat, and mammary gland ductal branching and that occurs at nearly 100% frequency in present day Native Americans and East Asians…was not fixed in USR1, Anzick-1, a Brazil_LapaDoSanto_9600BP individual and a Brazil_Laranjal_6700BP individual, all of whom carry the ancestral allele. Thus, the derived allele rose in frequency in parallel in both East Asians and in Native Americans.

These are on the older side as far as samples in the paper go. The numbers are small, but looking at modern Amerindian groups to have this much ancestral variant is surprising. The authors’ conclusion seems highly likely. The EDAR locus, and probably this particular SNP, was segregating in the ancient proto-East Asian/Amerindian metapopulation, and during the Holocene, there was selection on both sides of the Pacific.

Why? Unlike some people, I don’t think it was sexual selection for silky hair with full body. EDAR does a lot of things. From GeneCard:

The EDAR gene provides instructions for making a protein called the ectodysplasin A receptor. This protein is part of a signaling pathway that plays an important role in development before birth. Specifically, it is critical for interactions between two embryonic cell layers called the ectoderm and the mesoderm. In the early embryo, these cell layers form the basis for many of the body’s organs and tissues. Ectoderm-mesoderm interactions are essential for the formation of several structures that arise from the ectoderm, including the skin, hair, nails, teeth, and sweat glands.

This locus doesn’t seem to have been targeted elsewhere during the Holocene. Why not? Perhaps there’s another locus (or set of loci) that do similar things and were the targets of selection in other cases.

The bigger story, emphasized more in the other ancient DNA paper about South and North America that came out today in Science, Early human dispersals within the Americas, is that populations in the New World clearly seem to have been changing morphologically over the past ~10,000 years. Well, yeah….

* Ancient Mesolithic Scandinavian hunter-gatherers seem to have carried the derived variant at rs3827760. These people did not contribute much to the ancestry of later Scandinavians.