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

The deep origins of East Eurasians


The recent big paper on ancient DNA from East Asia has opened up a bit of a semantic can of worms. If you read all these ancient DNA papers with their stylized models you start to develop a sense of the big overall framework, but even in a big sprawling preprint with copious supplements, it is hard to make sense of things if you haven’t read what’s come before. With that in mind, I did a twitter thread where I outlined my own view and interpretation of how non-Africa was populated by modern humans in the last ~50,000 years with a focus on eastern Eurasia.

But I shouldn’t just leave it at Twitter. So I here I stand.

        click to enlarge

Above is a simple map, and to the right a stylized phylogram, that shows you the general gist of my thinking and what I’ve gleaned from the papers. First, we know that around 50,000 years ago there was a massive expansion from Africa or its margins all across Eurasia, and, that it reached Oceania really early.

Before that expansion it seems one group, we now call them “Basal Eurasians,” split off before the lineage that led to the peoples of Europe, Siberia, East Asia, South Asia, and Oceania. I set that at 60,000 years ago. Then around 50,000 years ago a group of populations that give rise to the peoples of Pleistocene East and South Asia, and Oceania, split off from groups that became the early Siberians (“Ancient North Eurasians”) and Pleistocene European hunter-gatherers.

I said Pleistocene because there has been massive population change across eastern Eurasia during the Holocene.

In South Asia about half the ancestry derives from populations with affinities or origins in western Eurasia, whether that be West Asia (Iranian-related farmers who occupied the northwestern fringe of the subcontinent expanding south and east), or further north in Central Eurasia and Eastern Europe (Sintashta-Andronovo). The balance is often termed “AASI”, or “Ancient Ancestral South Indian.” The mitochondrial evidence (lots of basal M) suggests deep and diversified lineages in South Asia, so I am willing to agree that this group descends from the first of the recent Out-of-Africa or Out-of-Africa-liminal* pulse of migrants 50,000 years or so ago. But, I do think it is not implausible that some of the ancestry of AASI derives from back-migration from Southeast Asia, which would be prime human habit during the dry and cold Pleistocene.

Further east, you have groups in Southeast Asia proper. If you listen to The Insight podcast you know that Pleistocene Southeast Asia was a much larger landmass due to lower sea levels. Not only was there more territory, but much of it was open wooded savannah, which often supports higher human populations than a tropical rainforest. In any case, most of the human population that lives in this region today descend mostly from farmers who occupied what is today the Yangzi river valley in China. There are exceptions. The Negritos of Malaysia and the Andamanese seem to be reflections of the peoples who were dominant in the region in the Pleistocene and most of the Holocene. Further east, there are Negritos in the Phillippines, who are distantly related to those further west but seem somewhat more connected to the Oceanians, even further eastward.

We know that the humans were in Australia by 40-45,000 years ago, at the latest. This establishes a timeline for the point of divergence of all these lineages. Though Andamanese are used as proxies for AASIs in population genetic analyses, it turns out they are very distantly related to them. All of these dark-skinned people across southern Eurasia and into Oceania are more related to each other than they are to East Asians, but only very distantly and marginally.

Speaking of which, the 40,000-year-old sample from Tianyuan near Beijing is the oldest representative of the human groups which we now term East and Southeast Asian. We know from this sample, and how it relates to other people in eastern Eurasia, that there was already significant differentiation across the region. I assume this had to have happened around 45-50,000 years ago. The time is less important then to note that the split between East Eurasians and West Eurasians, and within East Eurasians and West Eurasians, occurred in very rapid succession. This is the hallmark of an expanding species which occupies “empty” landscapes and fills all the possible niches very fast, and then stabilizes.

An easy dichotomy would be to label the Tianyuan people the northern clade, and what is often termed “Australasian” the southern clade. But it’s not that simple. The most recent paper, aligning with earlier results, argues that the Jomon people of ancient Japan (related to or ancestors of the Ainu) are about a 50:50 mix between the “northern” (Siberian?) and “southern” lineages. But Japan is in northern East Asia! Additionally, they also find that the oldest modern layer in Tibetans is also more closely related to the “southern” lineage. Finally, we know some populations in Amazonia are more closely related to the “southern” lineage than they “should” be, indicating that some “southern” ancestry came over Beringia.

To me, it makes it clear that the two lineages had a very different geographical distribution during the Pleistocene.

The Jomon culture dates to ~16,000 years ago, right after the Last Glacial Maximum.  We don’t know when the two distinct populations mixed, but I wouldn’t be surprised if it was quite early, as Siberian populations moved toward the coast. I assume that haplogroup D, found in both Tibet and Japan, has some origin among the “southern” people. Physical anthropologists have long noted some broad similarities in morphology between some “Australoid” people in India, Australian Aboriginals, and the Ainu of Japan. As the divergence between these groups is 45,000 years ago, the similarities may still be coincidences, but perhaps they are share ancestral characters?

We also need to think about Native American peoples. It looks like this group is a mixture of “northern” East Asians, related to the Devil’s Gate population, and ANE populations. This situates some of the post-Tianyuan groups in Siberia, but recall some Amazonians have “Australasian” affinities. The reality is that I think “southern” lineage population was long present along the Pacific fringe. The Jomon heritage is clear evidence of that. It is not implausible that there was structure in ancient Beringia, and some coastal populations with “southern” ancestry moved on earlier than the inland groups, and were mostly replaced except in the deep Amazon.

Finally, we can mention the strange Paleo-Tibetan “southern” ancestry. Again, there are peculiarities in Tibetan morphology which don’t make them in the same class as East Asians, but that can be attributed to high altitude adaptation. If the Paleo-Tibetans were the earliest population, they probably may have mixed with the Denisovans to obtain EPAS1. But, the majority of Tibetan ancestry seems to date to successive waves of “northern” populations that moved onto the plateau from the north and east. So who were the Paleo-Tibetans most closely related to? If I had to guess, I would say AASI. They may have moved onto the plateau from the south.

This is obviously exceedingly simple. There are many likely details I got wrong, as well as details I left out (I think that “southern” lineages were very common in the southern half of China deep into the Holocene, spanning the region between Tibetan and Jomon Japan). But, it gets across the gist of the broader framework in East Asia.

Here is the one-sentence version: rapid expansion, diversification, stabilization, then much more recent mixture between all the lineages.

* I believe that “modern” African humans were present in southern Eurasia in some form and numbers between 50 and 100 thousand years ago. But, I think they left a very light genetic imprint on modern populations. I accept that modern humans were in Sumatra 65,000 years ago, but I think almost all the ancestry is from groups which expanded 50,000 years ago, from the west.

30 thoughts on “The deep origins of East Eurasians

  1. @Razib – I think reading this in conjunction with Human Genome Diversity Project paper is particularly paper.

    Check out Figure 5 from the HGDP paper. Am I reading this correct? It looks to me that part of the Han/Papua split is on the order of 70-100kya. The same goes for the for the Han/West Eurasian split but to a lesser degree. The authors argue that this is likely real and not an artifact of the methodology.

    Is that evidence for a two-wave model?

    You said:

    “But, I do think it is not implausible that some of the ancestry of AASI derives from back-migration from Southeast Asia, which would be prime human habit during the dry and cold Pleistocene.”

    I think this is evidence for exactly that – an earlier peopling of SE Asia.

    I would add that your “clade 2” population is likely *not* the source of Denisovan admixture in East Asia. The HGDP paper makes it clear that Oceanians admixture from Denisovans that is separate and distinct from those experienced by East Asians, with the later coming from a source much closer to the Altai Denisovan population. This is consistent with the pattern in the Americas where the most Onge-like populations also had the least Denisovan admixture.

    Anyways, thoughts? I could flesh this out more.

  2. Looking at the Y-DNA, shouldn’t Ancient North Eurasians branch off much closer to Paleo-Siberians and Proto East Asians rather than Western Hunter Gatherers?

    For example Tianyuan man was Y haplogroup K, which is ancestral to both R and Q.

  3. Clade 2 which spread throughout South and Southeast Asia later replaced (not completely though) by farmers from West Asia and East Asia receptively. It shows impact of development of agriculture and shows agrarian societies could spread and dominate effectively more during pre-historic times. These societies later developed into more sophisticated civilizations (ancient Near East, neolithic and BA China) and continue to be even today more dominant.

  4. By the way, is it just me or is there there a Toba-shaped hole in the Denisovan ancestry in the world? The HGDP paper makes it clear that there are at least 3 sources of Denisovan admixture in modern populations – 1 in Oceanians (very distant to Altai Denisovans), 1 in East Asians (close to Altai Denisovans), and 1 specific to Cambodians (intermediate between Oceanian and Altai Denisovans).

    So everyone that moved north and east of the Toba ash cloud met their own unique Denisovan population but everyone in the footprint (or south or west) of the ash cloud didn’t meet Denisovans at all.

    At least that’s what it looks like from F7 in the HGDP paper.

  5. Previous paper also found that Jomon were already living in Japan by 38 kya, it would push back earlier than 16 kya from what i understand.

    “Therefore, our results support the archaeological evidence that the Jomon
    are direct descendants of the Upper-Paleolithic people who started living in the Japanese archipelago 38 kya.”

    https://www.biorxiv.org/content/10.1101/579177v1.full.pdf

    Now we have Mongolia_Neolithic along with Devils_Gate_Neolithic representing Siberian branch, bit too young in my view, any idea if older samples will be on the way? Considering East Asian and Siberian split is around 26 ybp according to the paper. Both Mongolia_N and Devils_Gate_N are directly ancestral to Native Americans, older samples will make things more clear.

  6. Razib,

    I’m leaving this comment here as the best way to contact you, so no need to publish it.

    Your podcast has recently starting mixing you hard to one side of the stereo channel, with the guest hard to the other side.

    Often, I have only one headphone in my ear when listening to podcasts (eg when dealing with check outs, crossing the road, trying to lie on my side to doze off). This means that I only get one side of the conversation.

    It would be possible to keep the stereo effect if you, say, mixed one side 70:30 and other other 30:70, rather than 100:0 and 0:100. It would leave me having to listen harder for one side of the conversation, but that’s manageable.

    No worries if this is more hassle than it’s worth.

    Thanks

  7. Good point by Ryan re the contrasting Denisovan vs. Population Y signals in Native Americans.

    Might I suggest you check out the long piece I recently posted on Academia.edu and Researchgate (“From Dyuktai to Clovis…)? I try to correlate archaeological and genomic data from NE Asia for Native American ancestry.

    On the Onge-like Population Y (from the paper):
    “There is, in fact, a very faint Denisovan signal in Native Americans (Qin and Stoneking 2015), as also in their distant East Asian relatives (but not in Europeans). These alleles are not the same as those of the mysterious Population Y, detected in the genomes of a few modern Amazonian natives (the Surui and Karitiana) (Skoglund et al. 2015). A previous finding of similar alleles in Aleuts (Raghavan et al. 2015) has not been replicated (Flegontov et al. 2019). Population Y seems to have been another “ghost”—a now-vanished East Asian population related to the Tianyuan Man, ca. 40,000 cal BP, from northern China, and to the remote ancestors shared by modern Andaman Islanders and Papuans (collectively, “Australasians”). The report of these alleles in a 10,400-year old skeleton from Lagoa Santa, Brazil (Moreno-Mayar et al. 2018), contrasts strangely with their absence from several 9600 year-old individuals from Lapa do Santo in the same region (Posth et al. 2018). These alleles are also absent from the much later pre-Columbian Taino of the Caribbean, whose ancestors came from the Orinoco basin (Schroeder et al. 2018). As Moreno-Mayar et al. (2018:4) acknowledge, “The fact that the Australasian genomic signature was present in Brazil 10.4 ka, but absent in all genomes tested to date as old or older and further north, presents a challenge in accounting for its presence in Lagoa Santa.” Posth et al. (2018) found not a single occurrence of these alleles in a sample of 49 ancient Central and South American genomes, which “raises the question of what ancient populations could have contributed the Population Y signal in Surui and other Amazonian groups and increases the previously small chance that this signal—despite the strong statistical evidence for it—was a false-positive.””

    The first archaeological evidence of humans in Amazonia is a light scatter of Clovis-derived Fishtail points across the Brazilian interior (I discount the amorphous broken rocks from NE Brazil). And the 9600 cal BP Lapa do Santo population in Brazil is genetically Anzick/Clovis descended.

    If the Population Y signal is not a “false positive,” shouldn’t it be appreciably stronger in the 10,400 cal BP Lagoa Santa sample than in modern Surui and Karitiana (who, apart from this oddity, are Anzick/Clovis descendants)?

    It would be very useful if the genome of some northern Japanese skeleton from ca. 15,000 cal BP (Incipient Jomon) were sequenced.

  8. @Junebug
    Tianyuan was K2b (pre R/Q), Ust’-Ishim and Oase were K2a (pre NO) and all of K2 has been proposed to originate from Southeast Asia in some papers, but Sikora et al. 2019 suggests ANE itself is a hybrid population of mostly early West Eurasian Sunghir/Kostenki14 related ancestry but there’s 20-30% northern East Eurasian (Devil’s Gate branch) ancestry too. This might just mean their Y-DNA is from the eastern side, early West Eurasian Y-DNA was mostly now rare branches of C.

  9. “we know some populations in Amazonia are more closely related to the “southern” lineage than they “should” be, indicating that some “southern” ancestry came over Beringia.”

    I think you are giving a bit too much big picture weight to this single data point of low percentage autosomal genetics from a very small number of tribes in a very isolated place in the far NW Amazon basin.

    For example, it is possible to model all of the existing Paleo-Asian ancestry in Amazonia as being due to a single family with that lineage that arrived in Beringia by boat in some fluke event a few years before the mass migration to the South along the Pacific Coast began and was part of a small tribe or band that stayed at the vanguard of that southern pacific coastal migration as a cohesive group until they followed a coastal river in South America up to its source at a continental divide and finally landed for good on the other side of that divide never to relocate again, without any introgression into any other New World populations.

    Something very much along those lines seems like the most parsimonious explanation, and it is hard to devise an alternative. Another would be a lone traveller or family of travelers whose boat is swept up in a storm from Oceania directly to the coast of Peru a few thousand years after the Founding expansion but pre-Columbian, who confused about directions migrates inland ending up with Founding population tribes and gets incorporated into them, tribes that may have been originally closer to the coast but migrated inland to form a relict population during the Inca or proto-Inca expansions.

    Significantly later migrations are problematic, because you don’t have a Paleo-Asian source population then.

    If there was an earlier Paleo-Asian migration, why do skeletal remains that seem archaic in the Americas have typical Native American founding population ancient DNA, and why aren’t their more obvious signs in terms of human remains and human made tools and structures and ecological impacts on American species in the pre-Founding population of the Americas out of Beringia era from a modern human population that was also expanding deeply into virgin territory?

    Any first wave migrants who weren’t on the vanguard of the wave of migration would have had their genomes admixed into Native Americans over a much wider swath of the Americas that half a dozen tribes deep in the Amazon after 20,000 years with even slight population exchange with neighboring populations.

  10. Tianyuan had K2b while modern east Asians have K2a. However, within K2a some of the most divergent subclades are found in south Asia and southeast Asia. On the other hand, within modern K2b, the closest branch to K2b2 (R + Q) is K2b1 which is a Papuan subclade. Did K2a and K2b change autosomal hands within ENA so to speak at some point?

  11. @DaThang

    The closest branch to RQ (or P1a) is a P clade found in the Phillipines (call it P1b). Another P clade more basal is found among Andamanese (P2). You can see these samples in YFull.

  12. @thejkhan

    Yes I know about those, should have clarified it as K2b2 (this includes R + Q).

  13. Check out Figure 5 from the HGDP paper. Am I reading this correct? It looks to me that part of the Han/Papua split is on the order of 70-100kya. The same goes for the for the Han/West Eurasian split but to a lesser degree. The authors argue that this is likely real and not an artifact of the methodology.

    are we sure this isn’t due to different archaic eurasian admixture?

    I would add that your “clade 2” population is likely *not* the source of Denisovan admixture in East Asia. The HGDP paper makes it clear that Oceanians admixture from Denisovans that is separate and distinct from those experienced by East Asians, with the later coming from a source much closer to the Altai Denisovan population. This is consistent with the pattern in the Americas where the most Onge-like populations also had the least Denisovan admixture.

    the oceanians and phillipine negritos have lots and lots of denisovan. the andamanese and malay negritos not so much. so yeah, totally diff.

    Looking at the Y-DNA, shouldn’t Ancient North Eurasians branch off much closer to Paleo-Siberians and Proto East Asians rather than Western Hunter Gatherers?

    For example Tianyuan man was Y haplogroup K, which is ancestral to both R and Q.

    the further in scale you look at Y and mtdna the less informative they’re gonna be. but yes, i lean toward the proposition that R & Q originally come from SE asia if you go back 50,000 years.

  14. are we sure this isn’t due to different archaic eurasian admixture?

    I suppose that’s a good point since the figure cuts off at 100kya, but it looks about ~10% of the genome doesn’t it? Seems a bit high given their other estimates for archaic contributions.

    the oceanians and phillipine negritos have lots and lots of denisovan. the andamanese and malay negritos not so much. so yeah, totally diff.

    That in and of itself could be consistent with encountering Denisovans part way along a coastal migration route. What I think is important in this paper is the finding that Oceanians and mainland East Asians received their Denisovan DNA from very different Denisovan populations.

    the further in scale you look at Y and mtdna the less informative they’re gonna be. but yes, i lean toward the proposition that R & Q originally come from SE asia if you go back 50,000 years.

    Y-DNA and mtDNA show very peculiar distributions going back to that depth of splits though. Like, why is mtDNA M virtually absent in modern West Eurasia? Why is the bulk of East Eurasian Y-DNA (K) nested in a primarily West Eurasian haplogroup (F), which itself is nested in seemingly East Eurasian CT.

  15. @Ryan

    Recent study suggests they are all from Southeast Asia and East Asia, where deep rooted lineages are. There are few more studies about this recently. This one comes to mind.

    “Early replacement of West Eurasian male Y chromosomes from the east
    We show that phylogenetic analyses of haplogroup C, D and FT sequences,
    including very rare deep-rooting lineages, together with phylogeographic analyses of ancient and present-day non-African Y-chromosomes, all point to East/South-east Asia as the origin 50,000-55,000 years ago of all known non-African male lineages (apart from recent migrants). This implies that the initial Y lineages in populations between Africa and eastern Asia have been entirely replaced by lineages from the east, contrasting with the expectations of the serial-founder model8,9, and thus informing and constraining models of the initial expansion.”

    https://www.biorxiv.org/content/10.1101/867317v1.full.pdf

    https://twitter.com/pille_hallast/status/1203043801855598593

  16. I’m curious about Basal Eurasians, what makes them distinct from other early Eurasians? is it drift along with lack of archaic admixture or is there more to it?

    “The statistic f4(Dzudzuana, X; Y, Mbuti) shows that early Eurasians like Ust’Ishim, Tianyuan and eastern non-Africans like Onge and Papuans share more alleles with European hunter-gatherers than with Dzudzuana”

    Dzudzuana is 26 ybp old sample with Basal Eurasian ancestry from Caucasus, they must have been more wide spread in Near East and Caucasus by 30 kyb?

  17. What if the “Australian” features of the Ainu/Jomon appearance is merely the presence of general features common to many ethnic groups? Having double eyelids, nose bridge, brows, facial hair etc. are not exclusive to any one group.

    I want to speculate (though I know it is impossible to be sure given what evidence I can find now) how much of these features would come from the clade 1 pre-Asian people they derive 55% of their ancestry from (and so do Native Americans 60-70% respectively), since some west coast native Americans (and Kennewick man) have been described Ainu-like in appearance.

    If typical “white” Europeans (blonde hair, blue eyes) did not fully make their appearance until not so long ago, then I doubt the typical modern east Asian appearance is any older.

  18. @Anthony

    Other 55% of Jomon ancestry is similar to indigenous Taiwanese, it’s not same as Native Americans.

    Indigenous Taiwanese have double-eyelid, along with those other features. It’s possible that some Jomon had light/blonde hair, since many native Australians do.

  19. Regarding these results, I assume that the Pleistocene Salkhit skull from Mongolia would be assigned to “Clade 1” of this analysis? Interesting that she has more West Eurasian ancestry than Tianyuan Man.

    “In recent years, the sequencing of many ancient early modern human genomes from West Eurasia has provided insights into the human population history in Europe [1]. In contrast, only four genomes from early modern humans in Siberia and one in East Asia have been generated, which limits our grasp of the genetic history of early East Eurasians [1]. Here, we present the genomic analyses of a hominin skull cap discovered in 2006 during mining operations in the Salkhit Valley, Northeastern Mongolia [2]. To our knowledge, the specimen remains the only Pleistocene hominin found so far in the country. Discovered outside any archeological context, the age and the ancestry of the specimen have been debated since its discovery and the presence of peculiar morphological features has led to potential affiliation to archaic hominin groups [3], [4]. We used a compound-specific radiocarbon dating approach to determine the age of the Salkhit individual to 34,950 – 33,900 Cal. BP (at 95% probability), placing the specimen in the Mongolian Early Upper Paleolithic period. We also determined its complete mitochondrial genome (mtDNA) and showed that it belongs to the modern human haplogroup N which is widespread in Eurasia today [5]. Nuclear analyses of the specimen show that the Salkhit individual was a female modern human. To investigate her relatedness to archaic hominin and ancient and present day modern humans, we used hybridization capture of 3.9 million single nucleotide polymorphisms across the nuclear genome. We use f3 and D statistics to show that she was closely related to the 40,000-year-old Tianyuan individual from China but shares more alleles with Western and North-Eastern Eurasians than does the Tianyuan individual. Using an admixture graph, the Salkhit individualis positioned as an ancient East Asian with some genetic contribution from West Eurasian. This scenario underlies an unexpected genetic link between East and West Eurasians after their major split. Using a new method to identify archaic introgressed DNA in ancient genome data, we estimate Neandertal ancestry in the Salkhit individual to∼2%, and show that this ancestry is contained in longer DNA tracks than those of present-day Eurasians – as expected given the age of the specimen. We also show that approximately 0.2-0.3% of the Salkhit genome is derived from Denisovans and identified longer tracts (>0.2 cM) of Denisovan ancestry in the Salkhit and other ancient East Eurasians genomes than in present-day East Eurasians genomes. This is the first evidence of Denisovan ancestry in Upper Pleistocene modern humans in East Eurasia. It is in sharp contrast to West Eurasia where we found no evidence for Denisovan introgression at the same resolution in early or later modern humans. The genome of the Salkhit individual provides further evidence for a complex population history of Pleistocene modern humans across Eurasia involving population substructure, migration and admixture. Those results emphasize the mosaic of events that shaped the genetic landscape of modern human since the Upper Pleistocene.”

  20. Tianyuan Man was K2b, I was trying to determine whether he was K2b* or K2b1* or K2b2* around this time last year but I abandoned the project and did not get back to it yet.
    He was positive on 2 of 4 SNPs defining K2b*, 2 no call. He was negative on 4 of 6 markers defining K2b1*, 2 no call. The list of K2b2* SNPs is very very long and it was extremely time consuming. He was negative on all stable called SNPs on that branch. So I devised a scheme to work with unstable SNPs but then I abandoned the project.

    My training was in theoretical math not comp sci so I don’t do much documentation when I write programs. I am a typical lazy mathematician too proud of myself. My program can be understood only by me and after some time even I find it hard to figure out what I was doing.

    My program I wrote last year was in BASIC initially because I thought that it would be enough but then it was not, so I refreshed on C++ programming and wrote programs in that language. And it was designed to work with only last year’s Reich lab data sets. Because I am lazy. It works only with a specific number of SNPs. It will take some time to fix it to work with this year’s data sets.

    Anyway according to my last year’s work, an Andamanese Man 100BC was the most divergent K2b2*. Several other men gave hints at being K2b2* but the SNPs were unstable(recurrent). All of them were on the mainland side of Asia(West of Wallace line) and all Oceanian K2b*s turned out to be K2b1.

    Next verified upstream K2b2 was the Yana man from Siberia ~ 32000BC. The Philipine man(not in the Reich data) was downstream to Yana but upstream to MA boy who was R*. To be precise the Philipino is not upstream obviously. He is a modern sample. But the common ancestor of him and MA boy lived somewhere between Yana and MA.

    On the other hand Yana either is or very close to the common ancestor of almost all living P’s including the Philipino but excluding the likes of the Andamanese 100BC if his descendant is alive somewhere. If Yana is not the direct ancestor, someone within about 4-6th cousins probably is because he harbored no extra SNPs. So I am not sure whether P originated in Southeast Asia. If it did it probably went up north and stayed there until it begot the MRCA of QR. Probably dwellling primarily among ANE.

    That is all I remember for now.

  21. @EastAsianMan

    Thanks. How could you tell Yana had no private SNPs, though? – that seems very unlikely a priori.

  22. I’m not really a pro on Y dna, but my understanding is that an issue with SNP capture on the Y is it will be missing anything not represented well in the modern pool that is used to design the target set.

    https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5418327/ – Kivisild 2017 – Several aspects of the capture-based data need to be considered when interpreting the phylogenetic reconstructions based on genotype rather than sequence data. First, the advantage of the capture approach, as mentioned above, is that it allows for the generation of data with higher coverage and higher overlaps among large number of individuals for a given set of SNPs (Pickrell and Reich 2014). This means that even in case of poorly preserved samples, robust haplogroup inferences, supported by multiple phylogenetically equivalent SNPs, can be made. The generic limitation of the approaches that focus on the enrichment-targeted SNPs is the ascertainment bias towards previously discovered variants and that the capture approach does not enable the user to discover novel variants and clades that have either become extinct or that have been not presented in the ascertainment set used for the design of the panel of SNPs to be captured (Fig. 2).

    If an ancient is from a line that just isn’t preserved at all, it may only present as being equivalent to the direct ancestor, but isn’t actually, really. And this is increasingly likely at a high time depth, but probably an issue even in relative recent periods (last 10kya) where expansions have reduced some clades down to low representation.

  23. @Megalophias
    The paper that presented the result for the Yana man explicitly said so.
    Reich’s set of YSNPs are very limited and I would not even attempt at such things.
    They attempted full sequencing much like with Ust Ishim but the quality of the sample was not as great. Even if it was of high quality the lack of private SNPs does not mean he is the direct ancestor but is close enough.
    In the Ust Ishim’s case they found several private YSNPs which gave the time estimate for the length of Ust Ishim’s private branch.(At the time it was not known that he was K2a, just K2(M526) so the number of private YSNPs were overestimated by 2 or so)

  24. @Cho

    Some native Americans on the west coast and Siberian groups like the Yukhagir have brown hair. If they’re from the same genes as Australian aboriginals, then I’d be very surprised. But I’m open to considering it if evidence comes out.

    That other 55% ancestry is clade 1, I didn’t say they were literally descended from the same group native Americans derive their ancestry from, but it’s a northern one.

  25. I just checked the Y-full tree and there, Yana is not ancestral to the Philipine sample. Last year there were some upstram SNPs of the Philipine sample that were ancestral in Yana. Maybe this year’s tree is correct but it also means that Yana probably is not ~32000 years old or the Yfull tree is messed up.
    SNP clocks can err but this is just too massive. MRCA of Yana and the Philipine sample was 30700 years ago according to the tree but Yana has 56 SNPs after that.

  26. Razib a proposed clarification to your ANE split in the diagram, it actually also shares a line with the East Asian (Northern) branch

    “Looking at the Y-DNA, shouldn’t Ancient North Eurasians branch off much closer to Paleo-Siberians and Proto East Asians rather than Western Hunter Gatherers?

    For example Tianyuan man was Y haplogroup K, which is ancestral to both R and Q.”

    yes ANE should tilt towards EEA, its not a fully independent branch, as its ancestor ANS (Yana/Ancient North Siberian) has 25% Tianyuan (EEA) ancestry:

    https://www.biorxiv.org/content/10.1101/448829v1.full.pdf

    “Using admixture graphs (qpGraph) and outgroup-based estimation of mixture proportions (qpAdm), we find that Yana can be modelled as EWE with ~25% contribution from EEA

    Among all ancient individuals, Yana shares the most genetic drift with Mal’ta, and
    f4 statistics show that Mal’ta shares more alleles with Yana than with EWE (e.g. f4(Mbuti,Mal’ta;Sunghir,Yana) = 0.0019, Z = 3.99; Extended Data Table 2; Supplementary Information 6). Mal’ta and Yana also exhibit a similar pattern of genetic affinities to both EWE and EEA, consistent with previous studies40,46 (Extended Data Fig. 3e). The ANE lineage can thus be considered a descendant of the ANS lineage, demonstrating that by 31.6 kya early representatives of this lineage were widespread across northern Eurasia, including far northeastern Siberia. ”

    Let me know if that makes sense

  27. Could you do a post on your thoughts on Basal Eurasian pls

    Great/informative East Eurasian posts so far

  28. Any idea how South Asians came to inherit ABCC11 allele? it’s frequency in South Asia looks to be connected with AASI.

  29. @Bmoney

    I am hoping the new findings will help clear up the nature of that, and where it belongs if confirmed, and whence.

    You’ll have to wait before Razib adds anything.

Comments are closed.