The long First Age of mankind

By Razib Khan | November 20, 2013 12:22 pm


“What it begins to suggest is that we’re looking at a ‘Lord of the Rings’-type world – that there were many hominid populations,” says Mark Thomas, an evolutionary geneticist at University College London who was at the meeting but was not involved in the work.

Mark Thomas, as reported by Nature

This is in reference to the ancient DNA meeting where David Reich reported that the Denisovans, an exotic archaic population which contributed ~5-10 percent of the ancestry of Papuans, was itself a synthesis of Neandertals and a mysterious group currently unknown. This is not surprising, as the broad outlines of these results were presented at ASHG 2012, though no doubt they’re moving closer to publication. But for this post I want to shift the focus to a different time and place, after the ancient admixture with archaic lineages, and to the reticulation present within our own.

But first we need to backtrack a bit. Let’s think about what we knew in the early 2000s. If you want a refresher, you might check our Spencer Wells’ The Journey of Man or Stephen Oppeneheimer’s Out of Eden, which focused on Y and mtDNA lineages respectively. These books were capstones to the era of uniparental phylogeographic analysis of the spread and diversification of anatomically modern African hominids ~50-100,000 years ago. Rather than looking at the whole genome (the technology was not there yet) these researchers focused on pieces of DNA passed down via direct maternal or paternal lineages, and reconstructed clean phylogenetic trees using a coalescent framework. Broadly speaking these trees were concordant, and told us that our lineage, all extant humans, derived from a small African population which flourished ~100,000 years ago. These insights suffused the thought of human evolutionary thinkers in other disciplines (see The Dawn of Human Culture). H. sapiens sapiens, veni, vidi, vici.

After that initial “Out of Africa” migration a series of bottlenecks and founder events led to the expansion of our lineage, as it replaced all predecessors. By the Last Glacial Maximum, ~20-25,000 years ago, the rough outlines of human genetic variation were established (with the exception of the expansion into the New World). We know now that this picture is very incomplete at the most innocuous, and highly misleading given the least charitable interpretation.

Reticulation. Graphs. Admixture. These words all point to the reality that rather than being the culmination of deep rooted regional populations which date back to the depths of the Pleistocene, most modern humans are recombinations of ancient lineages. On the grandest scale this is illustrated by the evidence of ‘archaic’ ancestry in modern humans. But even more pervasively we see evidence of widespread admixture between distinct lineages which are major world populations which we think of as archetypes. This is true for Amerindians, South Asians, and Europeans. This is also the case for Ethiopians, and Australian populations. A major problem crops up when we talk about extinct ancient populations which were the founding substituent elements of modern ones: it doesn’t make sense to use modern referents when they are simply recombinations of what they are describing. But language and history being what they weare we can’t change the awkwardness of talking about “Ancestral North Eurasians,” anodyne and somewhat incoherent at the same time (Eurasia is a modern construct with contemporary historical salience).

Into the mix comes another ancient DNA paper which reconstructs the genome of a boy who lived in Siberia, near Lake Baikal, somewhat over 20,000 years ago. It’s titled Upper Palaeolithic Siberian genome reveals dual ancestry of Native Americans. Here’s the topline finding: a substantial minority of the ancestry of modern Native Americans derives from a North Eurasian population which has closer affinities to West Eurasians than East Eurasians. And, this is an old admixture event. In the paper itself they observe that all “First American” populations seem to exhibit the same admixture distance to the Siberian genome. These results are also broadly consistent with the admixture of this population in Western Eurasia, especially northeast Europe. As among Amerindian populations it seems that this element is at substantial minority across Europe as a whole, and perhaps at parity in some populations, such as Finns.

Fig1To the left you see the geographical affinities of the MA-1 Siberian sample. It is shifted toward West Eurasians in the PCA. But on the map with circles representing populations, the definite evidence of admixture between Amerindians and MA-1 is clear in the shading. The statistic used, f-3, looks for complex population history between and outgroup (X) and a putative clade. From this test it is evident Amerindians had some admixture related to MA-1. Because of the dating of Siberian remains it does not seem likely that admixture was from Amerindians to West Eurasian and related populations. Rather, the reverse seems more plausible. You can also see from the map the close affinities with particular European and Central Asian populations of MA-1. This is intriguing, and requires further follow up. Though MA-1 and its kin were closer to West Eurasians than East Eurasians, it still seems likely that there was an early divergence between the populations of north-northeast Eurasia, and those of the southwest. Eventually they came back together in various proportions to produce modern Europeans, but it seems likely that during the Pleistocene these two groups went their own way.

treemixThere are hints of this in the TreeMix plot to the right. Note now drifted MA-1 is in relation to other West Eurasians (the branch is long). I suspect some of this is due to the fact that this individual is nearly 1,000 generations in the past. Not only is it difficult to name ancient populations with those of moderns, I suspect that some of the variation in the ancient populations has been lost, and so they seem exotic and difficult to fit into a broader phylogenetic framework (they had hundreds of thousands of SNPs though). And yet MA-1 can be fitted into the broader framework of populations which went north or west after leaving Africa because of mtDNA and Y chromosome results. Both of these indicate that MA-1 was basal to West Eurasians, with haplogroup U for mtDNA, and R for the Y lineage.

To really understand what’s going on here is going to take a while. A later subfossil, circa ~15,000 years before the present, yielded some genetic material, and exhibited continuity with MA-1. This suggests that Siberia may have had massive population replacement relatively recently. We know this was likely the case elsewhere. Reading Jean Manco’s Ancestral Journeys one possible scenario is that Pleistocene Europeans were MA-1 like, but were replaced by Middle Eastern farmers in the early Neolithic. But later eruptions from Central Asia brought mixed populations (Indo-Europeans?) with substantial MA-1 affinities to the center of European history.

Finally, one must make a note of phenotype. The authors looked at 124 pigmentation related SNPs (see supplemental). The conclusion seems to be that MA-1 was not highly de-pigmented, as is the case with most modern Northern Europeans. This stands to some reason, as substantial ancestry of this sort in Amerindians would result in phenotypic variation which does not seem to be present. Though the authors do suggest that coarse morphological variation among early First Americans (e.g., Kennewick Man) might be due to this population, which had West Eurasian affinities.

Where does this leave us? More questions of course. Though I’m confident the befuddlement will clear up in a few years….

Citation: doi:10.1038/nature12736

Addendum: Please read the supplements. They’re rich enough that you don’t need to read the letter if you don’t have access. Also, can we now finally bury the debate when east and west Eurasians diverged? Obviously it can’t have been that recent if a >20,000 year old individual had closer affinity to western populations.

CATEGORIZED UNDER: Anthroplogy, Archaeology
MORE ABOUT: Siberians

Comments (6)

  1. gpandatshang

    There are videos on Youtube of presentations at a conference starring Edward Vajda, whose work on Dene-Yeniseian is the sexiest thing to happen in linguistics in the last 25 years. Vajda’s talk pulls together linguistic, archaeological, and genetic evidence to create a hypothesis of how Siberian populations flowed into the New World. I know my linguistics quite a bit better than archaeology or genetics, and Vajda knows all three a lot better than I do; so I’m not in a position to assess his arguments critically, but he certainly sounds intriguing. The gist of his argument is that ca. 15kya, there was a population in south central Siberia that developed a micro-blade technology, which proved to be a “killer app” that allowed for rapid expansion. This group probably spoke Proto-Dene-Yeniseian (or, at least, the PDY speakers were early adopters). Male-only groups formed the leading edge of this expansion, absorbing existing female lineages in northeastern Siberia; these groups were the first to cross Beringia into the New World, bringing with them the non-Dene-Yeniseian languages spoken by most Indians. Meanwhile, the core micro-blade population, including males and females, expanded more gradually into the whole of Siberia. After thousands of years, one of these Dene-Yeniseian-speaking groups crossed the Pacific and settled what’s now the Pacific coast of Canada; this was the originally Athabaskan-Eyak-Tlingit population.

    In an interesting aside, Vajda argues against the idea (which I believe is the received opinion) that the first wave of settlers in North America consisted of a single, small group of hardy survivors who barely made it out of the Arctic before stumbling into Eden. He didn’t go into much detail on this point, but he apparently believes there were several somewhat larger groups who made the trip across Beringia at about the same time (long before the Dene-Yeniseian-speakers did). He suggests that the ancestors of today’s Algonquian-speaking groups were an intermediary population who were among the last of the first wave (i.e., they originated from a place in between the core micro-blade population and the northeastern Siberians who made up the bulk of the first wave).

    The key part of the presentation will come up on Youtube if you search for “Dene-Yensieian Workshop 2012, Edward Vajda”.

    I realize the above not directly on-topic as a response to OP, but it seemed to dovetail naturally.

    • razibkhan

      Vajda argues against the idea (which I believe is the received opinion) that the first wave of settlers in North America consisted of a single, small group of hardy survivors who barely made it out of the Arctic before stumbling into Eden.

      the genetics leans strongly against this. e.g. in this paper all ‘first americans’ from north to south america seem to be same distance from MA-1.

      but fascinating comment. this is one area i think that non-genetic fields can clarify and illuminate greatly.

  2. Karl Zimmerman

    I continue to find it interesting how well the MA-1 component maps to Northern European populations which have also been formerly identified as having higher levels of Mesolithic hunter-gatherer ancestry, and depressed levels of Neolithic ancestry. I wish they compared MA-1 with the known Mesolithic hunter-gatherer genomes from Spain and Sweden. It’s possible that “Ancient North Eurasians” were all European hunter-gatherers, and that modern Europeans are just Near Easterners + Ancient North Siberians.

    On the other hand, 20,000 years of racial continuity in an area which was not isolated geographically would be highly unusual. It’s highly plausible that Mesolithic Europeans were already somewhat modified not only by drift, but by earlier populations coming from the Near East and elsewhere.

    One thing I find interesting is the ADMIXTURE slide. MA-1 has the largest proportion of its ancestry (37%) as a South-Asian-like component. South Asians don’t rank particularly high compared to Northern Europeans when it comes to shared drift, but minus the highly ASI Paniya they’re well above Near Eastern populations (even Iranians). This seems to suggest that the Ancient North Indians themselves had a significant Ancient North Eurasian component. It was probably reinforced by later Indo-Aryan migrations however, given Indo-Aryans seem to have taken on Uralic loanwords after parting ways with Indo-Iranians.

    Regardless, one of the most fascinating papers this year.

    • razibkhan

      and that modern Europeans are just Near Easterners + Ancient North Siberians.

      good first approximation model IMO. probably more complex, but that’s a good sum.

      This seems to suggest that the Ancient North Indians themselves had a significant Ancient North Eurasian component.

      i wondered this too. look at the kalash in particular, as they don’t have east asian/siberian admixture (like the burusho). seems likely that PIE had fair amount of admixture of this + stuff coming up the caucasus from anatolia?

      • Karl Zimmerman

        I note that the level of shared drift for the Brahui and the Baloch is identical. To me this suggests that the vast majority of the MA-1 component came into the South Asian population prior to the Indo-European expansion. Yes other Dravidians tend to have much lower levels, but this is because they happen to have much more ASI.

        Also related to PEI, note that while Sardinians by far have the least shared drift, Basque are pretty typical of Southern Europeans, falling between Bulgarians and French. In addition, the Indo-European Armenians score lower than the non-IE North Caucasians, some of which (the Lezgians) have very high shared drift indeed. So there’s not much of a case to be made of a tight association with PIE and this component, although I’m sure some groups picked up substantive admixture along the way (Central Asia looks interesting – I wonder what it was like on this measure before the Turkic expansion.

      • Paul Conroy

        It seems that all my predictions are lining up.

        Mehrgarh for the likely origin of the R1b West European ancestry

        Southern Pakistan and West India as the likely home of Y-DNA P, which is the ancestor to R and Q


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar