Substack cometh, and lo it is good. (Pricing)

Eurasia became a melting pot during the Holocene


One of the things you notice when you look at genome-wide data are peculiar populations that seem to be shifted on PCA and other metrics in relation to exotic genetic affinities. For example, Sardinians, Japanese, and Taiwanese aborigines exhibit this pattern. When looking at Han Chinese data, many of the southern samples seem a bit further shifted away from West Eurasians than all the northern Chinese. That is, almost all northern Chinese seem to have low levels of West Eurasian affinity. Some of the southern Chinese do not.

When you look at West Eurasian data, you see evidence of East Eurasian gene flow into parts of Eastern Europe. Among Lithuanians, it seems to be there. It’s old and well-mixed, so it doesn’t jump out at you. But it’s there. Even more striking is that many of the Muslim populations in the Near East seem to have some proportion of East Asian ancestry because of the Turkic expansions.

We know the reason for this ancestry in West Asia. The rise of the Turks in the Islamic world is historically attested (thank you al-Mu’tasim!). Similarly, the arrival of Tatars and Magyars in Eastern Europe is also recorded. In China, various Turkic and West Asian populations arrived after the fall of the Han dynasty in the northern half of the country. I’ve documented on this weblog strong evidence of Indian ancestry across Southeast Asia.

As more ancient DNA comes to light I think one phenomenon that will become more clear is that the cultural tookit of humans over the last 10,000 years has allowed for more continuous, constant, and frequent, long-distance gene flow. Pairwise Fst values crashed with the rise of agriculture and larger-scale polities. But the adoption of the horse and the emergence of agro-pastoralism also served as a reciprocal conveyer belt of genes across the two antipodes of Eurasia.

West Eurasians and East Eurasians still remain genetically distinct. But evidence from Japanese and Sardinians gives a clear indication that within the last few thousand years have substantial reciprocal gene flows.*

* I am aware that in some of the work in David Reich’s lab there is evidence of East Eurasian gene flow into Mesolithic hunter-gatherers in Europe.

27 thoughts on “Eurasia became a melting pot during the Holocene

  1. One thing I’d be interested in your thoughts on is a pre-Holocene movement that doesn’t have as clear of an explanation as the domestication of the horse or the invention of farming.

    It’s the question of what drove ANE expansion and their replacement of the Gravettians in the Pontic-Caspian steppe, their presence in Northeastern Europe as EHG, their joining with Dzudzuana-origin farmers in sites like Ganj Dareh (with mostly R2a Y-DNA), and their eastern admixture to create the Native Americans.

    They’re a huge chunk of South Asian ancestry. After the demographic replacement/mixing of Central Asia with Turks & Mongols, it seems that some of the most ANE-shifted populations left in Eurasia are in South Asia (as well as the Caucasus), and while there is a North-South cline, a Velama from South India has roughly as much ANE admixture as a Northern European.

    But what’s more, this isn’t just ‘steppe’ contribution as it was in Europe. The IVCp population that makes up the ancestry of most of the subcontinent ranges between 28-43% which is higher than the most ANE-shifted modern populations in Europe. In contrast, the Sintashta (non-outlier) populations range from 36-44%, higher on average certainly, but not disproportionately so.

    The sheer range of this expansion seems interesting although I’m at a loss of what to attribute it to.

  2. The Japanese also exhibit some West Eurasian ancestry? If they do, from which source and when?

  3. roberto, they don’t. sorry if my prose was confusing. in contrast, a lot of han do. so that allows me to suggest that much of this is within the last 3000 years

  4. Interesting post, I am just confused about 2 points.

    Do you know how much West Eurasian gene flow exists in Han? For someone interested to read more about it, what would be some good papers?

    In addition, are you referring to Fu et al,2016 regarding that East Eurasian signal in the Mesolithic Hunter-gatherers of Europe? I think it is the paper that first described this affinity.

  5. Do you know how much West Eurasian gene flow exists in Han? For someone interested to read more about it, what would be some good papers?

    you can see evidence in charleston chiang group papers. i think it’s all over the place. i can run a pca for you and show you. it’s like 5% or so. probably mediated through mongols or turks (some han males have r1a).

  6. Thank you for the reply and the PCAs much appreciated!

    I found the paper ( A Comprehensive Map of Genetic Variation in the World’s Largest Ethnic Group-Han Chinese by Chiang et al,2018)

    Hopefully we will see some ancient DNA from China soon as well, it is a very understudied region.

  7. @surya: Most expansions were related to some genetic and/or cultural advantage of those being more successful. The ANE were not the only ones, expansions, mixtures and replacements happened all the time long before agriculture, horses and metals.
    The ANE were in any case physically quite impressive for their time, fairly tall, with large, robust skulls and body skeletons.

    But beside the possible advantages they might have had, even more important might be their motivation. And I think this relates to two events:
    1st: The Ice Age/LGM.
    2nd: The decline of the mammoth population/big game on the steppe.

    The ANE might have been the classic big game hunters and following quite often the mammoth herds.

    When the climate deteriorated and the game became less numerous, they seem to have left the Eurasian steppe in large numbers.
    So what we might observe are huge evasive movements of the steppe inhabitants.
    It would be interesting to know how the population density changed in this phase in different regions, but the data might be too scarce.

    My assumption is that they really emptied whole regions in this phase. When they moved down they could mix peacefully or conquer a better habitat.

    I’d say thats easier to research, because of uniparentals, yDNA lineages which point to the mode of admixture (win/paternal dominance, balanced/random or loss/female exogamy/no paternal impact).

    For example in Eastern Europe they were apparently totally dominant (uniparentals and autosomal admixture), not so in the Caucasus.

  8. Regarding the mesolithic mild east Eurasian affinity, could it be entirely explained by eastern WHGs having some ANE ancestry? Villabruna is like around 8% ANE and ANE is 25 to 32% ENA.

  9. @obs

    Do you have a link to a complete ANE specimen because I haven’t heard of one so far, just a few bones and fragments here and there. The only description I know of is the afontova gora 3 female being gracile.

  10. @DaThang: The decisive moment for the big West Eurasian : East Asian split was the LGM.
    There was little reason for a clear cut border along the Eurasian steppe before (or afterwards).

    Mongoloid proper is the direct result of the LGM adaptation in isolation of its core population from the Western Northerners. The LGM cut the steppe in two and the ANE hunters retreated, with one splinter moving to Beringia.

    Before that I would assume relatively constant small scale contacts and undifferentiated subpopulations in the middle which were neither WEA nor EA in the modern sense, but generally closer to WEA.

  11. @DaThang: You are right, there are only a few remains, but when the predominantly ANE derived EHG population appeared in the Mesolithic context, they looked exactly like that and the difference to WHG is quite striking.
    Even though Sunghir (adult male) is from a quite different (dead) branch, he lived under similar conditions and fits phenotype wise into the described trend.

    And in prehistoric as well as modern populations, there seems to be a clear trend for taller and more robust boned phenotypes with ANE ancestry.

    This trend is true in Europe (Sardinian-like vs. steppe-like) as well as in South Asia. In SA even beyond the IA steppe ancestry correlation, which is clear too in this respect, but not necessary (like Toda vs Panya).
    Of course, there were a lot of mixtures and selective trends in between, which could change and break the association completely, but generally speaking, the trend is real.

    It makes a lot of sense considering that they seem to have been among the best fed people of their time and had a lot of quite physical challenges in their habitat (big game hunting, intraspecific competition), plus the climate rules (cold adaptation by increasing size/total volume) and the availability of high Energie food resources.

    I don’t think EHG could have acquired their physique much later, they got it from their ANE steppe hunter ancestors.

  12. @Obs

    Thank you! Would be interesting to run sims to see if we could infer the relationship between carrying capacities of different ecological environments over time and patterns of admixture/replacement.

  13. @obs

    >@DaThang: The decisive moment for the big West Eurasian : East Asian split was the LGM.

    Yes- to extend it even further; but east Eurasian/ENA and west Eurasian were already showing up as different groups as far back as nearly 40,000 years ago (Tianyuan and Kostenki 14) though Ust Ishim from 45,000 years ago cannot be easily placed in either based on affinity alone, though a branch similar to him contributed more to certain later east Eurasians than to later west Eurasians. So it first began to form sometime between 45,000 and 40,000 years ago. Those 2 are obviously not as differentiated as modern east and west Eurasians are but the difference was emerging a pretty long time ago.

    >Mongoloid proper is the direct result of the LGM adaptation in isolation of its core population from the Western Northerners. The LGM cut the steppe in two and the ANE hunters retreated, with one splinter moving to Beringia.

    Even without the cold adaptation, the pre/proto-mongoloid ancestors of modern Mongoloids would end up being genetically/genotypically differentiated, but they would probably have a different, probably more tropically inclined phenotype.

    >Before that I would assume relatively constant small scale contacts and undifferentiated subpopulations in the middle which were neither WEA nor EA in the modern sense, but generally closer to WEA.

    While that is true, small scale interactions aren’t enough to fully bridge the gap and stop further differentiation. Remember- the middle east alone had 3 different populations before the neolithic mixing. One of the 3 (Iran) was quite different from the other two (Anatolian and Levant) and these didn’t begin to even out until the neolithic or the late epipaleolithic. Even without an LGM, while northern Asia would be a slow mixture of east and west Eurasians, the west Eurasians living in western Europe and east Eurasians in say China and Sundaland would experience very little if any trans-Eurasian geneflow. Exceptional scenarios would include a large otherwise unexpected movement due to events like the LGM or the end of the ice age or some crucial technological changes.

    >You are right, there are only a few remains, but when the predominantly ANE derived EHG population appeared in the Mesolithic context, they looked exactly like that and the difference to WHG is quite striking.

    EHG were kind of a mix: taller, but in certain cases more gracile than WHGs. There was this Sidelkino sample from Site 6, Burial 3, skeleton 1 who was genetically indicated to be a male but the phenotype had certain feminine indications. This was from the Mikulov Anthropology meeting 2014. I wouldn’t go so far as to claim that this group of traits came from ANE. I doubt that we can infer the ANE phenotype from EHG with a high degree of confidence. Intuitively, they should be robust due to mammoth hunting, though intuition could fail us, group of complete ANE adult skeletal samples are needed to arrive at a solid conclusion. So far I could only find description for an ANE female being gracile which is normal because she was a female.

    So for EHGs there was a period which could have allowed for Borealization, before the younger Dryas iirc, the summers were less hot and winters were less cool than today in the range in which they lived. Not saying that it is where their height or robustness/gracility comes from, but there are recent post-LGM periods allowing for changes over a few thousand years.

    Phenotype is subjected to change via biased/favoured selection- like mesolithic Iranians and south Asian AASI being both well built and relatively large bodied at once, these then resulted in smaller framed south Asians much later on.

  14. @DaThang: “Yes- to extend it even further; but east Eurasian/ENA and west Eurasian were already showing up as different groups as far back as nearly 40,000 years ago (Tianyuan and Kostenki 14) though Ust Ishim from 45,000 years ago cannot be easily placed in either based on affinity alone”

    Look, you can even have species at the extreme poles of a metapopulation with intermediate subspecies in between – so that gene flow can happen from one end to the other through the intermediary variants, but not from one end directly to the other if they would ever come into contact.

    So what we had before the LGM was a chain of subpopulations, with the extreme ends having little to do with each other, but still intermediaries and movements in between. This explains the minor, clearly from the opposite end, admixture here and there.
    ANE was phenotype wise quite clearly on the Proto-Caucasoid side of things, decisively so, but genetically they were the last piece of the chain which can be considered WEA.

    However, in the course of the LGM this chain was broken and most of the intermediary pieces were lost or pushed to the extreme fringes, with the end points evolving on, on their own. Especially the Mongoloid side at an accelerated pace on their very own, presumably because, unlike ANE, they couldn’t fully retreat from a habitat which became more extreme during the LGM.

    “Even without the cold adaptation, the pre/proto-mongoloid ancestors of modern Mongoloids would end up being genetically/genotypically differentiated, but they would probably have a different, probably more tropically inclined phenotype.”

    No, they were a winter season people even before. They would have been between undifferentiated Proto-Caucasoid, something on their own which does not exist now, came never into existence, or they would have gone largely the same path, just slower and not as extreme in some respects.
    If you want to know what kind of populations lived there, you have to look at the fringe groups of their sphere which survived, especially some Siberian and Indian people and the Ainu. They were not tropical (any more) at the time of the LGM at all. They were Northern and cold adapted before, just not as specialised, still more similar to the generalised West Eurasian spectrum.

    “While that is true, small scale interactions aren’t enough to fully bridge the gap and stop further differentiation. Remember- the middle east alone had 3 different populations before the neolithic mixing. One of the 3 (Iran) was quite different from the other two (Anatolian and Levant) and these didn’t begin to even out until the neolithic or the late epipaleolithic.”

    The Near East has significant barriers, but I wouldn’t agree with that assertion. Because the populations we see later most likely were already expansionist groups replacing others in parts of their later habitat. The picture is not complete now, we don’t even have more on Dzudzuana, which will be very, very important to have, still nothing about the “Basal Eurasians” and so on. There are gaps not just because of them being so important biologically, but because we lack the data (for now). The same is even more true for East Asia of course, were we know little to nothing about the Mongoloid evolution during the LGM, its a lot of speculation for the most part.

    “There was this Sidelkino sample from Site 6, Burial 3, skeleton 1 who was genetically indicated to be a male but the phenotype had certain feminine indications”

    Single specimen, especially if they show a strong deviation from the general norm, which might be even pathological, are never sufficient. But we have a lot of skeletal remains from later Mesolithic inhabitants which are more than 2/3 ANE and lived a similar life. And they show traits which are quite distinctive in comparison to the other people around them.

    “Phenotype is subjected to change via biased/favoured selection- like mesolithic Iranians and south Asian AASI being both well built and relatively large bodied at once, these then resulted in smaller framed south Asians much later on.”

    Well, the assumption is you deal with the same people, but you don’t, its not just selection on the same gene pool, its admixture and new trends associated with the population change too. Populations with a strong Iranian-like component are still overwhelmingly above average height and the Northern AASI population, which need to be better defined genetically first, doesn’t exist any more in an unmixed form. I doubt the AASI throughout South Asia were all the same in this respect as all.

    We can discuss these specific trends, but since its a theoretical debate, I say yes, generally its correct, because selective trends are key. Like it doesn’t matter how big the ancestors of the African Pygmies were, they show proof of an extreme selective trend for size reduction in the tropical forests.

    However, if a new physical trend appears with a new people and stays largely stable through time, the association makes sense and is in the ANE case for size justified.

  15. @Gonzalo

    EEF has been inferred to have Basal Eurasian admixture which would pull it away from East Eurasians I think.

  16. @obs

    >Look, you can even have species at the extreme poles of a metapopulation with intermediate subspecies in between – so that gene flow can happen from one end to the other through the intermediary variants, but not from one end directly to the other if they would ever come into contact.

    Yes, that is true. What I was trying to say is that the difference begins to show up already around 40,000 years ago from the previous undifferentiated population. A west Eurasian from Russia and an east Eurasian/ENA from China are found at the earliest around 40,000 years ago. This difference would be further magnified by isolation but it was already there. It wasn’t that big 40,000 years ago, about as much as the difference between Han and Hoabinhian.

    >So what we had before the LGM was a chain of subpopulations, with the extreme ends having little to do with each other, but still intermediaries and movements in between. This explains the minor, clearly from the opposite end, admixture here and there.

    I agree and this doesn’t necessarily conflict with a major difference beginning to emerge between 45,000 and 40,000 years ago. It wasn’t fully fledged or wide ranging as it became later on.

    >No, they were a winter season people even before. They would have been between undifferentiated Proto-Caucasoid, something on their own which does not exist now, came never into existence, or they would have gone largely the same path, just slower and not as extreme in some respects.

    The thing is, paleolithic skulls from China are closer to the Australo-Papuan range than the later post-LGM cold adapted ones.
    https://www.nature.com/articles/s41598-018-35426-z
    This isn’t like some sort of final declaration on the topic since the range of skulls from just before the LGM isn’t very extensive but the current evidence, even though it is weak, doesn’t support a pre-LGM cold adaptation.

    >If you want to know what kind of populations lived there, you have to look at the fringe groups of their sphere which survived, especially some Siberian and Indian people and the Ainu. They were not tropical (any more) at the time of the LGM at all. They were Northern and cold adapted before, just not as specialised, still more similar to the generalised West Eurasian spectrum.

    Almost all of them are still a byproduct of LGM adaptation- especially Siberians and native Americans. The Jomon skulls seem to fall in the Australo-Papuan/red range which is more tropical than the cold adapted populations (in the same link as before). There is also a shift in European HG cranial and skeletal traits due to the LGM.

    >The Near East has significant barriers, but I wouldn’t agree with that assertion. Because the populations we see later most likely were already expansionist groups replacing others in parts of their later habitat.

    The only major barrier between eastern Mediterranean and Iran was Zagros and the cultural technocomplexes on both sides did show some similar features (Kebaran and Zarzian), this could be due to some kind of communication. The populations might not have been entirely shut off but they were still differentiated regardless in that small region.

    >And they show traits which are quite distinctive in comparison to the other people around them.

    Can you provide some references for this? The only difference I know is that of average height and pigmentation.

    >Populations with a strong Iranian-like component are still overwhelmingly above average height and the Northern AASI population, which need to be better defined genetically first, doesn’t exist any more in an unmixed form. I doubt the AASI throughout South Asia were all the same in this respect as all.

    Today only Balochis and Bandaris are over 50% Iran descended and they aren’t particularly tall, just average. J Lawrence Angel in his Hotu cave paper notes how the neolithic and copper age Iranians themselves were quite different from the mesolithic Iranians of the Hotu cave in terms of being more massive.
    Agricultural selection was what I was hinting at- it selects against large HG traits which require a lot of calories and a diverse diet to sustain so the weaker and smaller individuals end up becoming more numerous in the presence of a specific agricultural diet. The effect needn’t be the same for all agricultural societies.
    As far as dealing with the same people or not is concerned- even if we look more recently, the Rakhigarhi skeletons were 175 cm on average for males and the steppe Sintashta average was also higher than the average south Asian height. Where did the short population input come from? I don’t think that unknown pygmies were involved in this, or even in the longer period of time from mesolithic to the present (when comparison mesolithic Iran HG, AASI to modern descendants). Even in neolithic Kashmir around 5000 years ago had individuals who were taller than the current inhabitants of Kashmir. An explanation could be that post-IVC farming in south Asia was probably not managed properly.

    >However, if a new physical trend appears with a new people and stays largely stable through time, the association makes sense and is in the ANE case for size justified.

    There is an association in terms of ANE ancestry and height among mesolithic European hunter gatherers on a skeletal level for sure, but that is all I can say based on what I know so far.

  17. @DaThang

    “What I was trying to say is that the difference begins to show up already around 40,000 years ago from the previous undifferentiated population. A west Eurasian from Russia and an east Eurasian/ENA from China are found at the earliest around 40,000 years ago. This difference would be further magnified by isolation but it was already there. It wasn’t that big 40,000 years ago, about as much as the difference between Han and Hoabinhian.”

    What exactly do you mean with the last sentence? Are you saying the genetic differentiation between Kostenki (I assume that’s the West Eurasian you’re referring to, or maybe Sunghir? I think they’re both around the same age, 35,000-40,000 ybp) and Tianyuan is comparable to the differentiation between Han and Hoabinhian? Sorry if I’m misinterpreting but that seems impossible. I was recently looking through Tianyuan paper and saw a few tables of comparative D-statistics plotting TY against East Asians and Europeans and he very strongly preferred the former over the latter in every instance. According to that paper, TY-like people didn’t directly contribute ancestry to modern Asians, but he was definitely part of a clade with them. Now there was indeed some sort of cryptic relationship between TY and Goyet, that is true, but not for other paleo-West Eurasians – if he’s an early Asian he should definitely be more distant from them than Han is Hoabinhian.

  18. @Mick

    I was basing this on raw global25 distances. I should have used the word distance as opposed to difference.

    BTW since you have the D stat info, do you know where he lands in relation to modern ENA populations or is he in his own branch?

  19. @DaThang

    You can check out the paper yourself here:

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6592271/

    There are a mess of different kinds of D stat tables in the supplements (Razib is 100% spot-on about supplements being more useful and interesting then the actual genetics papers themselves), I was mostly looking at Supplement 2 earlier so I really haven’t fully parsed it myself to be honest. IIRC, Tianyuan doesn’t seem to prefer one Asian over the other, however he does have a strange affinity to certain Native American tribes in South America over other Natives, and these Native Ams themselves also show an affinity to Papuan peoples. There is clearly some very interesting things going on with deep Asian population structure, hopefully we can get more Upper Paleothic from that part of the world soon.

  20. @DaThang: “The thing is, paleolithic skulls from China are closer to the Australo-Papuan range than the later post-LGM cold adapted ones.
    https://www.nature.com/articles/s41598-018-35426-z
    This isn’t like some sort of final declaration on the topic since the range of skulls from just before the LGM isn’t very extensive but the current evidence, even though it is weak, doesn’t support a pre-LGM cold adaptation.”

    It was known for a long time that Australo-Melanesian-like skulls and specimen roamed in Eurasia – in the West too. But the verdict is not final because we might deal with a layer which evolved on, or with a people which died out. This can only be elucidated by genetic testing and the proof of a direct genetic continuity. As long as that is not done, you don’t know with what you deal with.
    While at about 40.000 years ago the archaic/undifferentiated layer-model might be still closer to the truth, the same can’t be said for later times with the same certainty. Actually even in the early phase you find different variants which might be attributed to migrations from another source. Transient or not? Continuity or not? Especially about East Asia there is just a huge gap.

    But consider this from the abstract of the article you linked:
    “These two populations shared limited initial exchange, and the second layer grew at a faster rate and in greater numbers, linked with contexts of farming that may have supported increased population densities. Clear dichotomization between the two layers implies a temporally deep divergence of distinct migration routes for AMH through both southern and northern Eurasia.”

    That’s because they were different branches and only one was decisive for the Mongoloid/modern East Asian phenotype.

    “Can you provide some references for this? The only difference I know is that of average height and pigmentation.”

    They had very large craniofacial measurements and a very strong facial profile.

    With the rest you said I just agree with. But you might consider that a lot of tropical populations in South Asia and elsewhere are even more reduced than farmers in more favourable environments. The main reasons are energy (like you stated already) and intraspecific competition.

    ANE is even outside of Eurasia associated with fairly tall height. Most American Indians, unless they transitioned to a habitat and lifestyle which demanded something else, are fairly big. From North Western woodland Indians to South American Tierra del Fuego foragers.

  21. @Mick

    Thanks for the link.

    @obs

    I too agree with your post and yes there isn’t enough data yet so the results in that paper definitely aren’t conclusive. Its just that it is hard to define any hypothesis right now regarding the form of east Eurasian/ENA population dynamics as of now without further data, as Mick implies in his post. There was selection at some point, but I can’t find any information about a concrete period.

    IDK when exactly the different south Asian populations were reduced for each case (agriculture spread to different regions at different points of time), but the reduction was definitely considerable and there hasn’t been a ‘recovery’ yet.

    Lastly you do mention craniofacial ANE measurements. Do you have further specifications for that? I only have EHG cranial measurements and they don’t seem unusual in comparison to European HGs as far as basic, almost superficial comparison is concerned. However there is something that I read in a google book a while back (not the source of the EHG measurements that I have), something regarding small maximum frontal breadth and large minimum frontal breadth values.

Comments are closed.