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

Ancient humans had a lot of sex with each other

Have you taken my Steppelandia Quiz yet?

A new ‘must read’ paper on Neanderthals, Initial Upper Palaeolithic humans in Europe had recent Neanderthal ancestry:

Modern humans appeared in Europe by at least 45,000 years ago but the extent of their interactions with Neanderthals, who disappeared by about 40,000 years ago, and their relationship to the broader expansion of modern humans outside Africa are poorly understood. Here we present genome-wide data from three individuals dated to between 45,930 and 42,580 years ago from Bacho Kiro Cave, Bulgaria. They are the earliest Late Pleistocene modern humans known to have been recovered in Europe so far, and were found in association with an Initial Upper Palaeolithic artefact assemblage. Unlike two previously studied individuals of similar ages from Romania and Siberia who did not contribute detectably to later populations, these individuals are more closely related to present-day and ancient populations in East Asia and the Americas than to later west Eurasian populations. This indicates that they belonged to a modern human migration into Europe that was not previously known from the genetic record, and provides evidence that there was at least some continuity between the earliest modern humans in Europe and later people in Eurasia. Moreover, we find that all three individuals had Neanderthal ancestors a few generations back in their family history, confirming that the first European modern humans mixed with Neanderthals and suggesting that such mixing could have been common.

There are two primary points of interest:

  1. More evidence of ubiquitous ‘admixture’ between Neanderthals and ‘modern humans’ who may have been in contact with them.
  2. These earliest European modern humans in Southeast Europe seem to be more closely related to (possibly ancestral to?) people in East Asia than modern Europeans.

On the first point, twenty years ago there were some paleoanthropologists who were arguing there was no admixture between Neanderthals and modern humans. That is, no fertile offspring. To not pussyfoot around it, Neanderthals were barely acknowledged as human in the early years of this century by many scholars. This is evident in Richard Klein’s The Dawn of Human Culture, written in 2002. With hindsight much of this is ridiculous. It really didn’t make any sense that the Neanderthal and modern human lineage were not inter-fertile when you dug into the literature on mammalian hybridization. There just wasn’t that much time for them to be totally incompatible.

In 2010 when it was found that Neanderthal ancestry seems to be found in non-Africans, we updated many of our priors. I think it is clear on some level Neanderthal humanization is driven by the fact of Neanderthal admixture. Nevertheless, there was a plausible case this admixture was rare. Perhaps a single Neanderthal tribe was mixed into expanding modern humans? Perhaps a single Neanderthal? These were ideas that were mooted.

With what we know now I’m not sure this is tenable. There are two primary issues. First, there is variation in Neanderthal ancestry among non-African populations which does not seem to be due to African admixture. East Asians have more Neanderthal admixture (~25% more last I checked) than Europeans who have more than West Asians. South Asian Neanderthal admixture is proportional to their distance from East Asians. The less West Eurasian the South Asian, the more Neanderthal.

There are a few ways to understand this. A simple explanation is that there was a secondary admixture event that impacted people migrating to eastern Eurasia. Another hypothesis is that natural selection reduced the Neanderthal fraction over time, but East Eurasians were less impacted by this due to small effective population size. Then, there is the idea that a non-African population, “Basal Eurasians”, that did not have much Neanderthal admixture later mixed into West Asians and Europeans, reducing the original Neanderthal fraction.

Because of the small fractions and the paucity of ancient DNA more than 10,000 years old, there are still arguments around this. These results increase the probability that there were multiple Neanderthal admixture events. At least in my way of thinking.

The authors found that the Neanderthal fraction seems to have decreased from a higher level relatively early on. Rather than a gradual decrease, the authors found suggestions of strong effective selection in the first few generations. Though there is still room for the Basal Eurasian model, I’m not sure it’s quite as necessary now. Along with Oase, and the existence of a Neanderthal-Denisovan hybrid from Denisova cave, the likelihood of frequent admixture between ancient hominin lineages seems pretty high.

I believe the primary issue at this point is that the admixture event seems to be dated to a narrow time period, before the extinction of Neanderthals, but not too ancient (before 60,000 years ago), and the Neanderthal ancestry is quite similar. Perhaps there isn’t the power to detect multiple secondary admixtures at “around the same time.” From the perspective of today, 45,000 v. 50,000 years ago is “around the same time.” But the reality is 5,000 years is a long time.

A few days ago a collaborator of one of the authors above posted this preprint (last author), An extended admixture pulse model reveals the limits to the dating of Human-Neandertal introgression:

In simulations, we find that estimates of the mean time of admixture are largely robust to details in gene flow models. In contrast, the duration of the gene flow is much more difficult to recover, except under ideal circumstances where gene flow is recent or the exact recombination rate is known. We conclude that gene flow from Neandertals into modern humans could have happened over hundreds of generations. Ancient genomes from the time around the admixture event are thus likely required to resolve the question when, where, and for how long humans and Neandertals interacted.

The second major finding of this paper is that the very first modern humans to settle in Europe are more genetically close to East Asians than to modern Europeans. This is in contrast to the Oase sample from Romania, just to the north, and a few thousand years later, that was no closer to East or West Eurasians. Or, a new 45,000-year-old sample from Czech Republic, which also shows no connection to any modern people.

If you read the first paper, you know that years ago the GoyetQ116-1 sample from Belgium that dates to 35,000 years before the present, and which seems to have some ancestral connection to the later Magdalenian people of Pleistocene Europe, had an East Eurasian affinity. Additionally, earlier mtDNA work showed eastern affinities in some Pleistocene Europeans. The authors above make the connection between these eastern-affinity early Europeans and the “Initial Upper Paleolithic” (IUP), which seems to have expanded across a broad swath of Eurasian, from Europe all the way to western Mongolia.

The implication then is that some of the ancestry of East Eurasians comes from a migration that took a route through Europe, and around the Black Sea littoral. This is a bit more complicated than a pure “southern route”, but it’s not crazy, and has long been proposed.

I think at this point we need to take a step back, and acknowledge that the period between 40 and 60 thousand years ago is important, but we look through the glass darkly. Something happened, as all the ancient Eurasian hominin lineages were absorbed by the “Out of Africa” population (which may not have been expanding out of Africa!), and the phylogenetic relationships of some of these people do not make sense in light of the phylogeography of the present.

It is entirely feasible to me that the ancient non-African or proto-non-African populations had already started to develop some internal structure. The putative Basal Eurasian v. “the rest” is one key bifurcation, but there may already have been divisions between the western and eastern branches of Eurasians in northeast Africa or the Near East. We just don’t know yet. Basically, ancient substructure that is getting “blown up” with rapid radiation.

Finally, here’s the first author with a write-up, Ancient genomes and stone age encounters.

25 thoughts on “Ancient humans had a lot of sex with each other

  1. What would be the scenario where the “Out of Africa” population didn’t expand out of Africa? Is the idea that hominins were regularly migrating out of Africa well before ~50 kya, establishing multiple little population groups, and then one of those groups, living outside of Africa (but tracing its lineage back to Africa), rapidly expanded and displaced other extra-African populations…?

  2. It’s easy to visually juxtapose the Bacho Kiro Neanderthal segment chromosome maps with the continental-level ancestry segment maps in the African Americans or Latin Americans. The longest of the Bacho Kiro segments look the same, clearly a product of admixing which may have happened 5, 10, or 15 generations ago. But the vast majority of the segments are much, much shorter, leaving no doubt that they came from interbreeding on a scale of 100+ generations.

  3. IUP BK is closer to east Asians because it mixed with early differentiated east Eurasians to produce later east Eurasians. This distinction is based on Oase being crown Eurasian/outside of the proper ‘west Eurasian’ and proper ‘east Eurasian’ distinction. As I mentioned in the open thread, ‘east Eurasian’ and ‘west Eurasian’ without specifications are going to become less useful terms, even thought the branches that the terms normally refer to do exist.

  4. I would not go so far as to say the paper is biased but it is certainly very incomplete. What are the chances of Kostenki14 and Sunghir3 being relatively unadmixed – after the initial Neanderthal admixture – while all other populations go through a 3-ring circus of ancient admixture?

    “Basal Eurasians” was introduced to explain the relationship among modern populations and also among modern and ancient populations. It involves far more than just East Asians and Europeans. It remains to be seen whether this paper – even if it is right – obviates the need for Basal Eurasians. I doubt it.

    And in that process this paper also introduced 3 more Basal Eurasians. Basal 1 that contributed 3% to BK1653, Basal 2 contributing 98% to IUP Bacho Kiro and Basal 3 that contributed 83% to Ust Ishim.

    Even if this paper is right Basal Eurasians(even here at least 2 kinds) may still be needed to explain the relationship between Europeans, Near Easterners, ancient Iranians, South Asians, East Asians and Oceanians.

  5. The tree topology shows that Ust-Ishim should have a significant affinity to Tianyuan than to Kostenki14. This is incompatible with the known data.
    Likewise according to the tree topology, Ust-Ishim should be closer to East Asians than even to Kostenki14 and Sunghir3 (though an additional source for East Asian ancestry may neutralize it but how likely is it that the additional source exactly cancels it?).

    This paper may explain Ust-Ishim’s being closer to modern East Asians than to modern Europeans without the need for Basal Eurasians but in that process it fails to explain Ust-Ishim being equidistant with respect to East Asians and WHG.

  6. Sunghir isn’t unadmixed. This particular result paints it that way but more detailed results from the Denisovan paper of last year show that Sunghir is part Goyet as well, though not as much as main Gravettians were. Kostenki was kept on the side as the least admixed individual in that paper (like a reference), but it may have low level admixture instead of 0%. In order to know that you would need another individual less admixed than Kostenki to bring about the comparison.

  7. Re; Tianyuan, Ust-Ishim and Kostenki14; the paper states “This model uses 281,732 overlapping SNPs in all individuals and fits the data with a single outlier (Z = 3.22).”, and the supplement reveals that this outlier is relating to the relationship between Sunghir and Goyet relative to IUP Bacho-Kiro and Tianyuan: “Tia Sun Goy Bac -0.000258, Z: 3.219”. So this relates to the relationships between Sunghir and Goyet which DaThang describes above (where S and K are not exactly a clade relative to GQ116).

    As to the relationship between Ust-Ishim and T / K, it would seem that the excess of relatedness between UI and T implied by the model, relative to K, is below significance level (it’s less than Z=3). So isn’t as important to their model to fix as all the other things it gets right. Which are the greater relatedness of Goyet and Ust Ishim and Tianyuan to IUP BK than Kostenki has to IUP BK, and the greater relationships of GoyetQ116 to Tianyuan, Ust Ishim and IUP BK than Kostenki has.

    In a sense even a model with statistics which are all correct within significance levels is still imperfect, and a perfect model would fit the UI and T / K stat at zero (perhaps by adding some more admixture edges into K from UI or some other early splitting “ghost” UP population from Southeast Asia into T?). But it seems like this model is closer to being correct than other models, that might have admixture from T>IUP BK or GQ118>T or anything like this.

  8. @Medium,

    The typical alternative hypothesis to “Out of Africa” posits that the bulk of AMH ancestry was located in the Near East, or maybe North Africa. The reason why Sub-Saharan Africans appear in tree-based studies to split off first is they have considerably more archaic ancestry than Eurasians – but the bulk (greater than 50%) of their ancestry is from back-migrating Eurasians.

    We still lack the crucial data (significant amounts of ancient DNA from Africa) to prove or disprove the hypothesis. But the little bit we do have suggests the situation was far more complicated in Sub-Saharan Africa than we thought. For example, the Khoisan were traditionally seen as the most divergent/first group of humans to split off. However, while that might be true of the main branch of their ancestry, we know modern populations have considerable admixture with (likely Afro-Asiatic) East African pastoralists, who in turn were around 1/3rd West Eurasian. So they have not been isolated from Eurasian admixture even over the last 10,000 years. The same may hold true for other African populations. Y-haplogroup E in particular has always looked a bit odd, since it’s nested just within the “Eurasian” haplogroups, suggesting that it’s either a later Eurasian back-migration or that somehow its sister group (Haplogroup D) had a separate migration into Asia.

  9. Modern Europeans are more distant from Ust-Ishim than WHG is to Ust-Ishim and this paper does nothing to explain it. So Basal Eurasians are still needed, otherwise WHG needs to have an East Eurasian or Ust-Ishim admixture while modern Europeans are the more “pristine” West Eurasians.

    That sounds very convoluted in light of all the admixture events that led to the creation of modern Europeans. It violates the principle of parsimony and is a very typical amateur racist theory popular in some forums.

  10. Apparently I’m a Khan. I got a 22 on the quiz, though I don’t know what the maximum possible score is.

  11. @EastAsianMan, I don’t think the authors are suggesting that their model eliminates the need for Basal Eurasian* geneflow, and this is just their model of resolving relationships between the Upper Paleolithic samples.

    *or whatever we want to call it, and possibly North African or early Arabian Peninsula something.

  12. @Karl Zimmerman

    You wrote:

    “The typical alternative hypothesis to “Out of Africa” posits that the bulk of AMH ancestry was located in the Near East, or maybe North Africa…
    …The reason why Sub-Saharan Africans appear in tree-based studies to split off first is they have considerably more archaic ancestry than Eurasians – but the bulk (greater than 50%) of their ancestry is from back-migrating Eurasians.”

    One recent pre-print (Montinaro et al.) Does suggest as a possibility that many sub-Saharans may have a majority of their ancestry from an early (possibly “back-migrating”) group that lacked Neanderthal admixture, which either came from the Near East or somewhere in Africa (e.g. North Africa). But the study suggests that the “back-migrating” ancestry/population that Admixed with sub-Saharan Africans may have been from North Africa rather than Eurasia/Near East, which they say fits better with proposed Neanderthal/archaic admixture dates (“cautioning” that the population that admixed into Sub-Saharan Africans may not have been from outside Africa). They model the other native African ancestry in sub-Saharan Africans (abreviated as “AA”) as mainly basal modern/AMH, rather than archaic. The hypothesis of that paper would not seem to imply that AMH itself originated in the Near East or even necessarily North Africa, but rather that a certain ca. 60,000 ka branch may have (since AMH appear in Africa long before 60kya).

    Also, another reason sub-Saharan Africans appear to have split off earlier would be because they have ancestry from more basal AMH groups (populations that split off earlier than the ca. 60,000 ya divergence of the population ancestral to Eurasians) rather than necessarily (or only) because they have more archaic admixture (which has been suggested in West Africans for instance but not necessarily found in all sub-Saharans e.g. East African Nilotes/Nilo-Saharans).

    Another slightly earlier recent preprint (Cole et al.) Suggested that a smaller portion of the ancestry in many sub-Saharans could have come from a proto-Eurasian group that had not yet acquired Neanderthal admixture (around 30%, instead of the higher ammount suggested by Montinaro). But a recent review by Bergstrom, Stringer et al. (Referencing Cole in note 139) suggests that the back-migrating group detected in Cole et al could instead have been an African population, perhaps from North, East, or Northeast Africa (that had never left Africa but was related to the ancestors of Eurasians), similar to the alternate suggestion of Montinaro et al (that the source of the admixture could be from somewhere within Africa). They say:

    “An analysis of divergence times between segments of present-day genomes similarly suggested substantial admixture from a source related to the ancestors of non-Africans, but lacking Neanderthal admixture, into all African populations studied 139. This ancestry might never have left the African continent, but could represent an across-Africa expansion concurrent with the into-Eurasia expansion (Fig. 2a), and its spread could potentially be a major contributor to the complex genetic relationships observed among present-day African populations.” (page 5) (“Origins of modern human ancestry” by Bergstrom et al.)
    https://doi.org/10.1038/s41586-021-03244-5 (full paper here: http://генофонд.рф/wp-content/uploads/Skoglund-41586_2021_3244_OnlinePDF_300.pdf)

  13. @East Asian Man

    Do you think it’s possible that Bacho Kiro IUP and Tianyuan had some common ancestor from Central Asia?

    “That sounds very convoluted in light of all the admixture events that led to the creation of modern Europeans. It violates the principle of parsimony and is a very typical amateur racist theory popular in some forums.”

    What are you referring to? I don’t think this is what the authors are talking about, Basal Eurasians would a be a population splitting before Bacho Kiro.

    Also, I’d be interested in your thoughts of potential West Eurasian admixture in ancient East Asians making WHG closer to them than modern-day Europeans.

  14. Uniparental markers specific to Sub-Saharan Africa are not *that* divergent from non-African ones. If autosomal ancestry is more-or-less concordant with uniparental ancestry, then we’d expect layers of earlier diverging (but not ‘archaic’) ancestry to be important in Africa. There could also be genuine African archaic ancestry on top of that.

    Widespread African ancestry more closely related to Out-of-Africans – whether it originated inside or outside of Africa – is concordant with uniparentals, of course, since Y haplogroup E and the mt hg L3(xMN) branches are sister to or nested within Out-of-African variation.

    For African basal diversity be due primarily to archaics we’d need a huge whack of uniparental lineages to have back-migrated to Africa and died out outside of Africa, without bringing along attendant autosomal diversity. That does not seem parsimonious, to say the least.

    Has anything actually changed since this argument was debated at Dienekes’ blog 10 years ago?

  15. A clarifying edit to my previous comment:

    “[Many sub-Saharan groups seem more basal in large part because]…they have ancestry from more basal AMH groups (populations that split off earlier than the ca. 60,000 ya divergence of the population ancestral to Eurasians) rather than necessarily (or only) because they have more archaic admixture (some of which is suggested to be present in West Africans for instance but not necessarily in all sub-Saharans e.g. East African Nilotes/Nilo-Saharans – i.e. it is suggested that there is some archaic ancestry in West Africans but this is not the main reason they seem more basal than Eurasians; since West Africans, and most other Africans, also carry ancestry from basal or early-branching AMH/early modern groups).

  16. @James
    What is West Eurasians? According to Matt’s 2 layer theory there is no need for West Eurasian admixture(actually no need for West Eurasians in the first place). The admixture was suggested only by amateurs in some forums anyway.

    The differing proportion of Bacho Kiro-like and K14-like populations(more precisely a sister clade to K14) will result in a differing pseudo-West Eurasian affinity for East Asians.

  17. @East Asian Man

    ” According to Matt’s 2 layer theory there is no need for West Eurasian admixture(actually no need for West Eurasians in the first place).”

    What’s his theory?

    “The differing proportion of Bacho Kiro-like and K14-like populations(more precisely a sister clade to K14) will result in a differing pseudo-West Eurasian affinity for East Asians.”

    Kostenki is often classified as “West Eurasian” in the literature and certainly different from Tianyuan man in China. Do you think Tianyuan has K14-like ancestry or did that arrive later in East Asia.

    Basically I am kind of wondering what you think a likely model for East Asians (meaning Han-Japanese-Koreans here)

  18. @James The West Eurasian admixture signal is very faint and very geographically clined.(ie. Chinese having more of it than Koreans)
    So most likely it is due to the recent admixture.

    The shared drift of Tianyuan and WHG is very shallow but under the Bacho Kiro like + broader K14 like model(including Goyet and Tianyuan part2) proposed by Matt, the more of the latter will result in a higher pseudo West Eurasian affinity – albeit faintly.

    Matt however seems to think that Tianyuan harbors a genuine K14 component and it is the same one as in Yana. I am not even sure Yana has it. Yana has a West Eurasian component but have not seen anyone claiming it is specifically Kostenki-14 related outside of amateur circles.

  19. @East Asian Man

    Interesting, would you know how much of it we are talking in Han? I knew some Xibei populations have it (something like 5% in certain cases) but not present across all China for example (above trace levels) First time I hear it about Koreans tbh.

    “(including Goyet and Tianyuan part2) proposed by Matt, the more of the latter will result in a higher pseudo West Eurasian affinity – albeit faintly.”

    So if I understand you correctly, Tianyuan seems to act as a vector for that affinity. I do remember reading Fu et al 2017 showing Tianyuan affinity with a subset of WHGs, I think they implied the migration was from east to west though.

    K14 keeps getting referred to as “West Eurasian” even in papers, hence why I called it that. If I recall correctly it’s also found in ancient populations in Europe.

  20. Another 45,000 YBP sample Zlatý kůň is also closer to Asians. Despite the sample being pre-West-and-East split.

    https://www.nature.com/articles/s41559-021-01443-x

    “We first compared Zlatý kůň with present-day European and Asian individuals using an African population (Mbuti) as an outgroup and found that Zlatý kůň shares more alleles with Asians than with Europeans (Extended Data Fig. 6). A closer relationship to Asians has also been observed for other Upper Palaeolithic and Mesolithic European hunter-gatherers compared with present-day Europeans and can be explained by ancestry in present-day Europeans from a deeply divergent out-of-Africa lineage referred to as basal Eurasian.”

    “When we tested European hunter-gatherers without basal Eurasian ancestry against ancient and present-day Asians, we found that none of these comparisons indicate a closer relationship of Zlatý kůň with either group (Supplementary Sections 5 and 9 and Extended Data Fig. 7). This suggests that Zlatý kůň falls basal to the split of the European and Asian populations.”

    This all begs the question about where did “West Eurasians” actually come from..Near East? North Africa?

  21. @James

    In interview author says : “Hajdinjak suggests that the Bacho Kiro remains represent a population that once lived across Eurasia, but vanished from Europe and lived on in Asia.”

  22. @Cho

    “In interview author says : “Hajdinjak suggests that the Bacho Kiro remains represent a population that once lived across Eurasia, but vanished from Europe and lived on in Asia.”

    Right my question is whether that pop spread to both Europe and East Asia from Siberia/Central Asia.

Comments are closed.