The deep origins of East African Hunter-Gatherers

PNAs has a new paper out, Genomic evidence for shared common ancestry of East African hunting-gathering populations and insights into local adaptation. From what I can tell this was never a preprint, so it’s all new….

Or is it? Looking closely at some of the populations sampled, I’m about 85% sure that I saw a very early and preliminary analysis of some of these data (probably on a different SNP-chip) at ASHG 2012. I say this because I recall talking to the second author in front of the poster about an obscure hunter-gatherer tribe in Ethiopia that they had sampled. Unlike some graduate students he did not dodge my inquiries by standing away from the poster as if he was not associated with it!

Here is the abstract:

Anatomically modern humans arose in Africa ∼300,000 years ago, but the demographic and adaptive histories of African populations are not well-characterized. Here, we have generated a genome-wide dataset from 840 Africans, residing in western, eastern, southern, and northern Africa, belonging to 50 ethnicities, and speaking languages belonging to four language families. In addition to agriculturalists and pastoralists, our study includes 16 populations that practice, or until recently have practiced, a hunting-gathering (HG) lifestyle. We observe that genetic structure in Africa is broadly correlated not only with geography, but to a lesser extent, with linguistic affiliation and subsistence strategy. Four East African HG (EHG) populations that are geographically distant from each other show evidence of common ancestry: the Hadza and Sandawe in Tanzania, who speak languages with clicks classified as Khoisan; the Dahalo in Kenya, whose language has remnant clicks; and the Sabue in Ethiopia, who speak an unclassified language. Additionally, we observed common ancestry between central African rainforest HGs and southern African San, the latter of whom speak languages with clicks classified as Khoisan. With the exception of the EHG, central African rainforest HGs, and San, other HG groups in Africa appear genetically similar to neighboring agriculturalist or pastoralist populations….

Some of this stuff was vaguely predictable a long time ago. There is a strange tendency in older data and results for hunter-gatherers such as Pygmies and San Bushmen to be closer together genetically against agro-pastoralists and farmers. Additionally, the two most deeply diverged Y chromosomal haplogroups, A and B, tend to be found in African hunter-gatherers in particular. At least at high frequencies.

The main phylogenetic result of this work is some other isolated hunter-gatherers in eastern Africa, more obscure than the Pygmies, Hadza, and San Bushmen, also seem to show deep affinities that set them apart from demographically dominant groups such as Nilotic pastoralists and Bantu farmers.

This is not surprising though in light of ancient DNA. A few years back Pontus Skoglund’s paper showed that there was likely a preexistent relatedness cline in East Africa between the peoples who were present in Ethiopia before the arrival of Eurasians and south toward the ancestors of the modern Khoisan groups in southern Africa.

This work likely confirms the existence of this cline before it was demographically swamped out by migrations triggered by groups which had shifted their modes of production (from hunting and gathering to farming and/or pastoralism). That being said, I am somewhat skeptical of very deep history inferences from these data. In the paper itself, the authors admit that the origins of modern humans in Africa date to ~300,000 years before the present. Skoglund’s data set is from the Holocene, the last ~10,000 years or so.

The authors used a Bayesian framework to make a few concrete inferences:

The maximum a posteriori estimate and 95% credible interval for pairwise divergence time estimates are shown in Fig. 3. The maximum a posteriori divergence time estimates for the Hadza and Sandawe were 13 or 22 kya when accounting for different primary sources of admixture based on STRUCTURE analysis (Fig. 1D) (NC or AA, respectively); these estimates overlap with previous studies (17). The Hadza split times with other populations were older; the divergence time estimates with the Sabue (NC or AA gene flow) were 44 or 61 kya, respectively, and with the Dahalo (NC or AA gene flow) were 55 or 61 kya, respectively. Sandawe population divergence time estimates with the Sabue (NS or AA admixture) were 30 or 52 kya, respectively, and with the Dahalo (NS or AA gene flow) were 50 or 57 kya, respectively. The estimated times of divergence of the Sabue and Dahalo (NC or AA gene flow) were 63 or 72 kya, respectively. These results are consistent with a model in which population divergence between the Dahalo and Sabue and the ancestors of the Sandawe and Hadza occurred >30 kya, whereas the Hadza and Sandawe divergence was more recent. Whole-genome sequence analyses will be informative for more accurately resolving the time of population divergence among EHGs.

There are a few things I’ll note about this. Some of these populations, such as the Sandawe, are clearly the product of recent admixture. This is supposedly accounted for, but who knows what details remained unaccounted. Additionally, a lot of the divergence times are weirdly coincident with the “Out of Africa” migration, ~50,000 years ago. Finally, I have zero confidence that hunter-gatherers in Tanzania and Ethiopia had ancestors in the local region ~50,000 years ago. They may have. But we can’t take for granted that these were static groups. There’s strong evidence that Pygmies have undergone selection for smaller size, so we know that they weren’t in the deep rainforest for hundreds of thousands of years.

Rather than specifics, what I take away from this is that there is a lot of deep structure in East Africa which was swamped out by recent demographic expansion by Bantu farmers. We knew this. But there is a lot more detail here that needs to be explored.

One thing I want to note is that the most common Y chromosomal lineage within Africa is a branch of E1, which diversified ~50,000 years ago. This is nearly fixed in many agriculturalist groups. The Yoruba are over 90% E1b1a. But even the Hadza are 40% E1. Some of this is likely from admixture with Bantus, but some of it is probably from admixture with Cushitic people. There have long been arguments about the origins of this haplogroup. It turns out that the ancient Natufians carry it.

But, it is highly possible that E is indicative of a serious “back migration” into deeper Africa coincident with the “Out of Africa” movement (is origin may have been in North Africa as opposed to West Asia). Several years ago David Reich mentioned offhand that Eurasians may be closest to the Hadza of unadmixed African groups. This is not geographically implausible, but what if the patterns that Reich’s group sees are due to an ancient back migration from the Middle East deep in the Pleistocene? One thing ancient DNA is telling is that many “unmixed” populations do have admixture.

The pattern of deep and ancient structure within Africa is only vaguely understood at this point. Models which infer parameter estimates based on current understandings may give misleading inferences because we are specifying a model that misses major features of the true dynamics. For example, there is evidence for a deeply basal lineage in West Africa, as well as continuous gene flow from West Eurasians well after the “Out of Africa” event.

As they say. “Watch this space” and “developing”….