The mixed genetic origin of the first farmers of Europe

[DgidguBidgu]

Nobody said all Q is related to PIE. Some subclasses of Q1a though yes. There's also plenty of non IE R1b like M73 and and V88 and some old subclades of R1a as well.

Some are clearly confusing PIE with the IE group of languages ​​today. PIE refers to a limited group, the first and only native speaker of the original Indo-European language. Today's languages ​​included in the IE group belong to it solely due to the fact that they share a set of words occupied by the PIE and nothing more. These are different languages ​​by origin, different in structure, grammar and age, visible even today. They have literally nothing to do with PIE, neither in the past nor today and this is a constant.
So I personally deeply doubt that even some R1a line has a direct connection with PIE. They were simply Indo-Europeanized quite late. And Q has nothing to do with the topic of IE languages. Pure isolates dragged from certain source migrations. The PIE is thought to have had contacts with Asia, so there is nothing inexplicable in its sporadic appearance in Europe.

[TuaMan]

Here is the bottom line:

[LIST][*]Eurasians are characterized by the Eurasian bottleneck characterized by the "Main Eurasian" node here

The "OoA" bottleneck was a contributing factor to Eurasian genetic structure but I don't think it's right to attribute all or even most of the Eurasian vs SSA genetic variance to the bottlenecking that seems to have occurred around 50-60 kya years ago. "Eurasians" are not some SSA Africans that got bottlenecked 50-60kya years ago, they were distinct from what would become anything resembling SSA long before the bottleneck. Multiple papers have dated the divergence between Eurasians and SSA to upwards of 100 kya or more. Some examples:

https://www.biorxiv.org/content/10.1...06.01.127555v1

https://genomebiology.biomedcentral....059-019-1684-5

Africa is genetically very diverse and not some sort of genetic monolith. Some components seemed closer phylogenetically to Eurasians than to each other by far[*]But what all "SSA" (whatever you wish to call it) components share is that they clearly did not partake in the Eurasian bottleneck and they cluster off to the other side

Some other people in this thread have already discussed the common node (y-DNA DE/E and mtDNA L3) linking Eurasians and Africans; what I think defines SSA as SSA are the non-CT, non-L3 lineages that all SSA have (not necessarily exactly in common) that Eurasians do not. I think all SSAs, whether they're Niger-Congoid, Nilo-Saharan, Pygmy HGs, San Bushmen derive at least 50% of their ancestry from non-CT, non-L3 lineages, and that side of their ancestry is what makes SSAs, SSA. Because ultimately those are the only lineages we can be 100% sure originated in Africa; if we take seriously the multiple estimates of Eurasian vs African divergence going back to 100 kya than by definition that puts CT and L3 more on the Eurasian side than it does the African. Because if Eurasians started separating from their nearest African kin that long ago, what other ancestral uniparental lineages could they possibly have other than CT and L3?

I also think people exaggerate the great internal diversity of Africans - yes, the HG groups like San and Pygmies definitely seem pretty diverged from all other human groups, but the overwhelming majority of Africans are not HGs but either Niger-Congoid or Nilo-Saharan speakers and these two groups don't seem to be radically diverged from each other. This study dates the divergence between NC and NS speakers to as recently as 28 kya for example:

https://genomebiology.biomedcentral....059-019-1679-2

That said, nothing about internal African phylogeny is really that clear right now. I think it's mix of not having enough aDNA from Africa and probably the current SNPs panels that are standard not really being optimized to capture African variation.

Are you suggesting that E-M35 entered the region with their Dzudzuana-related side? That doesn't make much sense. I think it seems likely that the pre-ANE & North-African admixture Y-DNA profile of West-Asia were lineages like G, J, T and L, no? E-M35 in these guys is likely owed to their ANA ancestry which then syncs fairly well with their autosomes. ANA Y-DNA and Eurasian mtDNA. I think ANAs probably had a lot to do with Y-DNA E. Can't be a coincidence that Mota who seems more closely related to whatever ANA might be than other "SSA" components was E-Y175024 which is a subclade of E-P177 just like E-M215, the ancestor of E-M35. And the split between E and D happened a long time ago. Around 65,000ybp which may very well fit with how, in terms of auDNA, ANA seems close to Eurasians compared to other non-Eurasian components. Am I missing something here?

I didn't mean to insinute that E-M35 was owed to the Dzudzuana side of IBM's ancestry, I was just trying to articulate that I found the connection of ANA to some diverged 70 kya+ population to be suspect. I think it's definitely fair to link E-M35 to ANA, but I think if you do that it doesn't make sense to carry on that IBM is really around 50% ANA for the reasons I outlined in my most recent response to Keneki20, whatever the formal stats might say. I think there's a "there" with ANA, but I don't think it's really as high in IBM as the current models suggest.

[Mnemonics]

Mnemonics, if you don't mind one more request, does Morocco_EN (IAM) need additional Levantine input in order to be successfully modeled in qpAdm? Or do they basically just look like unadmixed IBM descendants?

They don't seem to require anything more than Iberomaurusian as the single source.

[VladimirTaraskin]

Some are clearly confusing PIE with the IE group of languages ​​today. PIE refers to a limited group, the first and only native speaker of the original Indo-European language. Today's languages ​​included in the IE group belong to it solely due to the fact that they share a set of words occupied by the PIE and nothing more. These are different languages ​​by origin, different in structure, grammar and age, visible even today. They have literally nothing to do with PIE, neither in the past nor today and this is a constant.
So I personally deeply doubt that even some R1a line has a direct connection with PIE. They were simply Indo-Europeanized quite late. And Q has nothing to do with the topic of IE languages. Pure isolates dragged from certain source migrations. The PIE is thought to have had contacts with Asia, so there is nothing inexplicable in its sporadic appearance in Europe.

After such statements, I think you need to submit an application for the Nobel prize in comparative linguistics. Allan Bomhard is silently envious.

[Ryukendo]

The "OoA" bottleneck was a contributing factor to Eurasian genetic structure but I don't think it's right to attribute all or even most of the Eurasian vs SSA genetic variance to the bottlenecking that seems to have occurred around 50-60 kya years ago. "Eurasians" are not some SSA Africans that got bottlenecked 50-60kya years ago, they were distinct from what would become anything resembling SSA long before the bottleneck. Multiple papers have dated the divergence between Eurasians and SSA to upwards of 100 kya or more. Some examples:

https://www.biorxiv.org/content/10.1...06.01.127555v1

https://genomebiology.biomedcentral....059-019-1684-5




Some other people in this thread have already discussed the common node (y-DNA DE/E and mtDNA L3) linking Eurasians and Africans; what I think defines SSA as SSA are the non-CT, non-L3 lineages that all SSA have (not necessarily exactly in common) that Eurasians do not. I think all SSAs, whether they're Niger-Congoid, Nilo-Saharan, Pygmy HGs, San Bushmen derive at least 50% of their ancestry from non-CT, non-L3 lineages, and that side of their ancestry is what makes SSAs, SSA. Because ultimately those are the only lineages we can be 100% sure originated in Africa; if we take seriously the multiple estimates of Eurasian vs African divergence going back to 100 kya than by definition that puts CT and L3 more on the Eurasian side than it does the African. Because if Eurasians started separating from their nearest African kin that long ago, what other ancestral uniparental lineages could they possibly have other than CT and L3?

I also think people exaggerate the great internal diversity of Africans - yes, the HG groups like San and Pygmies definitely seem pretty diverged from all other human groups, but the overwhelming majority of Africans are not HGs but either Niger-Congoid or Nilo-Saharan speakers and these two groups don't seem to be radically diverged from each other. This study dates the divergence between NC and NS speakers to as recently as 28 kya for example:

https://genomebiology.biomedcentral....059-019-1679-2

That said, nothing about internal African phylogeny is really that clear right now. I think it's mix of not having enough aDNA from Africa and probably the current SNPs panels that are standard not really being optimized to capture African variation.



I didn't mean to insinute that E-M35 was owed to the Dzudzuana side of IBM's ancestry, I was just trying to articulate that I found the connection of ANA to some diverged 70 kya+ population to be suspect. I think it's definitely fair to link E-M35 to ANA, but I think if you do that it doesn't make sense to carry on that IBM is really around 50% ANA for the reasons I outlined in my most recent response to Keneki20, whatever the formal stats might say. I think there's a "there" with ANA, but I don't think it's really as high in IBM as the current models suggest.

I think a big reason why the split times are so different between Schlebusch et al and Fan et al for intra "Neo-African" splits is because of different populations they use: the assumptions cause the split times to be scaled up or down (Schlebusch et al ~250 kya vs Fan et al ~100 kya for the San split), but more importantly the populations they pick are not the same, Fan et al picked Rendille for the Afroasiatic group and Sengwer for the Nilo-Saharan group, and if you look at e.g. Scheinfeldt et al Rendille and Sengwer are quite similar to each other despite speaking different languages and neither have very much "Nilo-Saharan/East African" ancestry compared to the levels in Dinka, which would produce underestimates.

[rothaer]

What about discussing the paper?

Placing Loschbour and Bichon ("west 1, "west 2") in Western Europa at the LGM seems wrong as this WHG (Villabruna cluster) first showed up in the northern Adriatic at abt 17 kybp and the Western parts of Europe at that time were inhabited by Magdalenians with GoyetQ2-like DNA.

[Riverman]

I also think people exaggerate the great internal diversity of Africans - yes, the HG groups like San and Pygmies definitely seem pretty diverged from all other human groups, but the overwhelming majority of Africans are not HGs but either Niger-Congoid or Nilo-Saharan speakers and these two groups don't seem to be radically diverged from each other. This study dates the divergence between NC and NS speakers to as recently as 28 kya for example:

https://genomebiology.biomedcentral....059-019-1679-2

That said, nothing about internal African phylogeny is really that clear right now. I think it's mix of not having enough aDNA from Africa and probably the current SNPs panels that are standard not really being optimized to capture African variation.

Exactly, Eurasians are different for many reasons, including admixture, later drift and selection - as well as being descendants of a specific subset of Africans, a heritage shared with the ANA-like population element. There was no average African profile from which they stem, because it didn't exist at that time, since there were different regional populations, probably at some point forming a meta-population, but a lot is conjectural without hard evidence for anything.

In the article you linked they too speak about Niger-Congo/Kordofanian in terms like may or may not, both the timing and the place of origin is unsure, as the Shum Laka paper showed, which just proved some mainstream theories wrong or at least wasn't able to prove them. They can only provide a time window and tell us that it was a recent expansions, but that's about it.
Another issue I have with branching events estimates, because most start with the premise that there was one group, then they split, that's it. But in reality, on the ground in Africa, its highly likely that there was some gene flow and admixture events taking place, which could bring the estimates for the original (non-complete) split back by quite a lot.
Especially groups like Niger-Kordofanians were no unity, they might not even have existed, as a modern-relatives like, genetic population, just 10.000 years ago. They are the result of fairly recent population dynamic, and that's about the most important thing the Shum Laka paper seems to suggest.

That's like if we looked at the start of admixture runs and statistical analyses of the autosomal DNA at Northern Europeans and recognised that there was some kind of affinity with Caucasians, more so than to ancestral Southern Europeans. Now we know why, because of "steppe ancestry" being the result of a mixture event. For the core West African group, which spread with Niger-Kordofanian ethnicities, its about the same thing. We can now see that there are some affinities and these being caused by admixture events, but the details not being worked out yet.

[Awale]

The "OoA" bottleneck was a contributing factor to Eurasian genetic structure but I don't think it's right to attribute all or even most of the Eurasian vs SSA genetic variance to the bottlenecking that seems to have occurred around 50-60 kya years ago. "Eurasians" are not some SSA Africans that got bottlenecked 50-60kya years ago, they were distinct from what would become anything resembling SSA long before the bottleneck. Multiple papers have dated the divergence between Eurasians and SSA to upwards of 100 kya or more.

I'm well aware but I do think the main contributing factor to the PCA divide among Eurasians and Africans is in fact the bottleneck. It's the only drift event that really makes sense. If it was in fact just normal differentiation over-time I don't think it would specifically group Eurasians off to the side together like that the same way it doesn't do that for the SAHGs and whatnot as they are so far proposed not to have shared ancestry with some other African groups even longer than Eurasians have.

Some other people in this thread have already discussed the common node (y-DNA DE/E and mtDNA L3) linking Eurasians and Africans; what I think defines SSA as SSA are the non-CT, non-L3 lineages that all SSA have (not necessarily exactly in common) that Eurasians do not. I think all SSAs, whether they're Niger-Congoid, Nilo-Saharan, Pygmy HGs, San Bushmen derive at least 50% of their ancestry from non-CT, non-L3 lineages, and that side of their ancestry is what makes SSAs, SSA. Because ultimately those are the only lineages we can be 100% sure originated in Africa; if we take seriously the multiple estimates of Eurasian vs African divergence going back to 100 kya than by definition that puts CT and L3 more on the Eurasian side than it does the African. Because if Eurasians started separating from their nearest African kin that long ago, what other ancestral uniparental lineages could they possibly have other than CT and L3?

Personally I don't think there was really a clear autosomal "Eurasian" & "African" divide going that far back (L3 and E). I maybe wrong but that's what I suspect. This would have been a time when Eurasians and their closest relatives in Africa or the Middle-East may have, in terms of time-divergence, been split from some other African lineages for quite some time but the stark divide you see in something like the global PCA or ADMIXTURE likely had not formed yet and I suspect the "closest relatives" group did not participate in whatever caused this stark "Eurasians Vs. Everyone else" divide while still in terms of phylogeny remaining close relatives of the Proto-Eurasians and carrying lineages like L3 and E. Whether or not they back-migrated or pretty much never left Africa as some are arguing over is hardly important in my opinion.

I also think people exaggerate the great internal diversity of Africans - yes, the HG groups like San and Pygmies definitely seem pretty diverged from all other human groups, but the overwhelming majority of Africans are not HGs but either Niger-Congoid or Nilo-Saharan speakers and these two groups don't seem to be radically diverged from each other. This study dates the divergence between NC and NS speakers to as recently as 28 kya for example

Yes, and people are pretty much talking about the overall scope of "Africans" when they talk about internal diversity. The continent was relatively recently homogenized in part by things like the Bantu expansion, the expansions of NS speakers and AA speakers and so on. And some of these groups' deeper ancestries like between NC and NS speakers seem relatively mixed into each other to boot. Some of the stuff Keneki20 and others pointed out about L3 lineages among West-Africans weren't lost on people even about a decade ago. I remember Lank pointing this sort of thing out to me as far back as 8 years ago or so and I think even that is a late date for when people noticed it. The mtDNA has pointed for a long-time to something related to the dominant ancestry in East-Africans being present in West-African NG speakers like the Yoruba too. But if you go back to Africa before some of these expansions and earlier intermixtures I don't think it is at all an exaggeration to note the internal diversity as groups like SAHGs, central HGs and whatnot would've had a wider spread and even some wide regions all to their themselves then have seemingly intermediate regions with theoretically "AEA" rich groups like Mota going through the Horn and SE Africa then if ANA is a real construct whatever it is would have been somewhere up north of all this and already mixed with a Dzudzuana-like group and I haven't even gotten to what really early West-Africa and some other parts might have looked like and we'll have to see from more ancient DNA. But as you can see, pre-recent expansions Africa had some pretty deep divergences and diversity going. Much more than what is found in Eurasia. And even when you look at the post expansion groups, stuff like the mtDNA and even Y-DNA diversity to some extent looks a lot crazier than anything in Eurasia.

I didn't mean to insinute that E-M35 was owed to the Dzudzuana side of IBM's ancestry, I was just trying to articulate that I found the connection of ANA to some diverged 70 kya+ population to be suspect. I think it's definitely fair to link E-M35 to ANA, but I think if you do that it doesn't make sense to carry on that IBM is really around 50% ANA for the reasons I outlined in my most recent response to Keneki20, whatever the formal stats might say. I think there's a "there" with ANA, but I don't think it's really as high in IBM as the current models suggest.

I'm stating the obvious but we really do need more aDNA from Africa and the very early Middle East to properly discuss any of this. For now, I don't think whatever "ANA" is can be denied as having clear substantial SSA-related affinities. Consider that every paper IBMs have been shown in pretty much models them as significantly SSA-related in some way even if the models of how that works differ then you have all the non-academic runs failing to find any real alternative as well. The PCA position of IBMs seeming to sync with this is just an aside, frankly. But what I think ANA might be if it's real may very well be something related in part to the modern "AEA" cluster we've all thrown around for years and in turn a fair amount of the non-SAHG-like ancestry in Mota too. I suspect things like the extensive degree of L3 lineages found among groups who have a lot of such ancestry and Y-DNA E might be pointing to this. There was most likely a group in Africa with lots of L3 and E who were also in terms of autosomes phylogenetically very close to the Proto-Eurasians but did not take part in their bottleneck somewhere in North & East Africa or the Middle-East or both in the far ancient past (hardly matters where) and a lot of the ancestry in African groups like Yorubas, Dinkas & even ancients like Mota most likely descends from this group as may ANA if it is a real component. I think this accounts for how these groups show so much E & L3 as well as well phylogenetic shifts toward Eurasians over some other African groups like SAHG despite most of their ancestry appearing "Non-Eurasian".

I have seen this sort of scenario proposed for over a decade. Except in the past before all this aDNA the focus was on something "East-African" as these autosomal affinities toward Eurasians peaked in groups like Dinkas when you excluded significantly Eurasian admixed groups like Cushites, Ethiosemites & Omotic speakers. I think the ANA construct or at least some part of it might be this component finally being seen in action in some way outside of the obvious admixtures in groups like Dinkas, Horners, Mota, West-Africans and so forth. And I'm not just talking about Yoruba-like admixture in groups like the Dinkas when I refer to them as admixed away from this ancient theoretical group as you might notice through their uniparentals.

What about discussing the paper?

Placing Loschbour and Bichon ("west 1, "west 2") in Western Europa at the LGM seems wrong as this WHG (Villabruna cluster) first showed up in the northern Adriatic at abt 17 kybp and the Western parts of Europe at that time were inhabited by Magdalenians with GoyetQ2-like DNA.

Yeah, I think I'll reply to Riverman over PMs out of respect but this last post to TuaMan is probably it for me. We've derailed the thread enough.

[Riverman]

I'm well aware but I do think the main contributing factor to the PCA divide among Eurasians and Africans is in fact the bottleneck. It's the only drift event that really makes sense. If it was in fact just normal differentiation over-time I don't think it would specifically group Eurasians off to the side together like that the same way it doesn't do that for the SAHGs and whatnot as they are so far proposed not to have shared ancestry with some other African groups even longer than Eurasians have.

Two factors of importance to add to the proposed bottleneck:
- All later African components might have intermixed at least to some degree, while the Eurasians descended exclusively from one group, from ANA-like ancestry also got at least most of its own, but probably not all, if they admixed in the course of the back migration - just as a possible scenario.
- Eurasians have increased Neandertal admixture. So while Proto-ANA got pulled in the direction of Basal Africans, Proto-Eurasians got pulled in the direction of Neandertals.

How much gene flow, seletion and earlier drift played in, we will see.

Personally I don't think there was really a clear autosomal "Eurasian" & "African" divide going that far back (L3 and E).

I think that's largely a question of geography and terminology, but from the genetic point of view there was the derived, fully modern H.s. group from which both Proto-ANA and Proto-Eurasians branched off. Where that happened, is open to debate, but most likely are North East Africa and/or the Near East. The big split happened when E (or at least non-E1b1b) plus L3 branches migrated into the rest of Africa and spread in North Africa and beyond. The next time Eurasians and Africans met was probably with IBM. So there would be a time window of about 30.000-50.000 years in which they were probably separated. In this time Proto-ANA spread in Africa and mixed with locals, while Proto-Eurasians split in their respective branches and migrated throughout the continent and beyond. Like two brothers one moving West, the other East, founding their own families, going their own ways.

But if you go back to Africa before some of these expansions and earlier intermixtures I don't think it is at all an exaggeration to note the internal diversity as groups like SAHGs, central HGs and whatnot would've had a wider spread and even some wide regions all to their themselves then have seemingly intermediate regions with theoretically "AEA" rich groups like Mota going through the Horn and SE Africa then if ANA is a real construct whatever it is would have been somewhere up north of all this and already mixed with a Dzudzuana-like group and I haven't even gotten to what really early West-Africa and some other parts might have looked like and we'll have to see from more ancient DNA. But as you can see, pre-recent expansions Africa had some pretty deep divergences and diversity going. Much more than what is found in Eurasia. And even when you look at the post expansion groups, stuff like the mtDNA and even Y-DNA diversity to some extent looks a lot crazier than anything in Eurasia.

I agree with what you said in this paragraph by and large, but there is one problem. You talk about Mota as if this is of real relevance, but I tell you, that's like looking at an LBK sample from Europe and talking about "Paleolithic European variation", it won't work out. Mota is, most likely, a fairly new immigrant group, for the most part, its not ancient Eastern African, like in Europe something like Neandertals and Magdalenian, not even WHG. Which brings us kind of back to the original topic

I don't think whatever "ANA" is can be denied as having clear substantial SSA-related affinities. Consider that every paper it's been shown in pretty much models it as significantly SSA-related in some way

If you have a child born form a union of father from A and mother from B, the child will have, in any case, affinities to A. Otherwise it wouldn't be the child. But if you separate the fathers genome, from that of the mother, he can be much closer to anything but the mother. Father A can have children with any female which is fertile, that doesn't tell you anything about his affinities and position in the tree. If you look at it from the perspective of the child ("SSA"), surely, ANA is the father. But that doesn't make it closer to the mother's ancestry necessarily. And ANA was only slightly closer to other African branches, than Proto-Eurasians. And the main reason is a later bottleneck, selection and a higher degree of Neandertal admixture on the one, but Basal African ancestry on the other side. So from the moment ANA would have migrated back and mixed with locals, they would be much more different. Considering that, they were still fairly close on a general human tree at the time of IBM, which is kind of remarkable.
Earlier finds and their ancestors in North East Africa or the Near East might be still even closer. Who knows? Probably that idea is wrong, but it doesn't look like that to me right now.

even if the models of how that works differ then you have all the non-academic runs failing to find any real alternative as well. The PCA position of IBMs seeming to sync with this is just an aside, frankly. But what I think ANA might be if it's real may very well be something related in part to the modern "AEA" cluster we've all thrown around for years and in turn a fair amount of the non-SAHG-like ancestry in Mota too. I suspect things like the extensive degree of L3 lineages found among groups who have a lot of such ancestry and Y-DNA E might be pointing to this. There was most likely a group in Africa with lots of L3 and E who were also in terms of autosomes phylogenetically very close to the Proto-Eurasians but did not take part in their bottleneck somewhere in North & East Africa or the Middle-East or both in the far ancient past (hardly matters where) and a lot of the ancestry in African groups like Yorubas, Dinkas & even ancients like Mota most likely descends from this group as may ANA if it is a real component. I think this accounts for how these groups show so much E & L3 as well as well phylogenetic shifts toward Eurasians over some other African groups like SAHG despite most of their ancestry appearing "Non-Eurasian".

Exactly. Seems we can agree largely on this. Question is just where it happened and what caused the ANA - Eurasian stark genetic separation. Which is something none can answer without more samples and better methods, new papers and data.

Another view back to the original topic, because after all those papers and data coming in, there are still debates about the exact origin and character of the original farmers. This just points to the long way to go for Africa, which is far from anything like that concerning testing volume.

[beyoku]

You talk about Mota as if this is of real relevance, but I tell you, that's like looking at an LBK sample from Europe and talking about "Paleolithic European variation", it won't work out. Mota is, most likely, a fairly new immigrant group, for the most part, its not ancient Eastern African, like in Europe something like Neandertals and Magdalenian, not even WHG. Which brings us kind of back to the original topic
.

Fairly new from where exactly? And how is "New" defined in age?