Zeng et al: Postglacial genomes from foragers across Northern Eurasia...

[Pribislav]

Postglacial genomes from foragers across Northern Eurasia reveal prehistoric mobility associated with the spread of the Uralic and Yeniseian languages

https://www.biorxiv.org/content/10.1101/...1.560332v1

Abstract
The North Eurasian forest and forest-steppe zones have sustained millennia of sociocultural connections among northern peoples. We present genome-wide ancient DNA data for 181 individuals from this region spanning the Mesolithic, Neolithic and Bronze Age. We find that Early to Mid-Holocene hunter-gatherer populations from across the southern forest and forest-steppes of Northern Eurasia can be characterized by a continuous gradient of ancestry that remained stable for millennia, ranging from fully West Eurasian in the Baltic region to fully East Asian in the Transbaikal region. In contrast, cotemporaneous groups in far Northeast Siberia were genetically distinct, retaining high levels of continuity from a population that was the primary source of ancestry for Native Americans. By the mid-Holocene, admixture between this early Northeastern Siberian population and groups from Inland East Asia and the Amur River Basin produced two distinctive populations in eastern Siberia that played an important role in the genetic formation of later people. Ancestry from the first population, Cis-Baikal Late Neolithic-Bronze Age (Cisbaikal_LNBA), is found substantially only among Yeniseian-speaking groups and those known to have admixed with them. Ancestry from the second, Yakutian Late Neolithic-Bronze Age (Yakutia_LNBA), is strongly associated with present-day Uralic speakers. We show how Yakutia_LNBA ancestry spread from an east Siberian origin ~4.5kya, along with subclades of Y-chromosome haplogroup N occurring at high frequencies among present-day Uralic speakers, into Western and Central Siberia in communities associated with Seima-Turbino metallurgy: a suite of advanced bronze casting techniques that spread explosively across an enormous region of Northern Eurasia ~4.0kya. However, the ancestry of the 16 Seima-Turbino-period individuals--the first reported from sites with this metallurgy--was otherwise extraordinarily diverse, with partial descent from Indo-Iranian-speaking pastoralists and multiple hunter-gatherer populations from widely separated regions of Eurasia. Our results provide support for theories suggesting that early Uralic speakers at the beginning of their westward dispersal where involved in the expansion of Seima-Turbino metallurgical traditions, and suggests that both cultural transmission and migration were important in the spread of Seima-Turbino material culture.


Fantastic paper, can't wait for the BAMs to be released! And if I'm not wrong, the lead author is one of AG blokes most of us are very familiar with, so hats off to him.

[RCO]

Have you found any file/table with the new Y-DNA samples ? I could only find old published Y-DNA samples.

[Andar]

Most northern ancient R1a-Z93 found yet and it is from Seimo-Turbino (Site 6 on the map). The dating is probably incorrect and i guess there is reservoir effect and should rather be around 2000 B.C but still very interesting. Also seems to be rich in Yakutia_LNBA and WSHG admix.

I32552 2571-2348 calBCE Russia_SeymoTurbinoCulture Satyga-16 (Khanty-Mansi Autonomous Okrug, Kondinskiy District, Yagodny) Russia R-Z2122 R1a1a1b2a2

I32552. In the first set of qpAdm, this individual can be modeled as three-way admixtures between~47% Yakutia_LNBA + ~5-20% Srubnaya_LNBA, with the other two sources drawn from the FSHG cline (p>0.44). These models are in keeping with the behavior of this individual in ADMIXTURE (Fig.4B, Extended Data Figure 10). In the second set of qpAdms, with populations of the Eneolithic/Early Bronze Age Trans-Ural Steppe as sources, 3-way admixture models pass (p>0.10), with ~40% each of Russia_Tatarka_BA-related and FSHG-related ancestry, and ~15-20% ancestry from the Steppe_MLBA-related source. All passing models for this individual include a Yakutia_LNBA-related source.

[Anglesqueville]

I agree with pribislav, it's a fantastic paper. We have not finished talking about it, in particular for its implications on the location of the Proto-Uralic. I hope it doesn't take them years to publish the data.

[PopGenist82]

I like it - covers facts, context & theory.

[Zelto]

Have you found any file/table with the new Y-DNA samples ? I could only find old published Y-DNA samples.

The new samples are mixed with the old in Table 2 of the Data S1 supplementary file. Go to column "F" and sort by "This publication".


Unrelated:

Has anyone manually checked the Y-calls of I1961? He is that previously published twist sequence sample we now have context for, originally claimed to be N-L1026. However, I saw someone report that he is actually N-Y9023; a subclade also common in Uralic speakers. This result would make a lot more sense as N-L1026 (N-Z1979 as a whole) are intimately associated with Yakutia_LNBA admixture and initially formed/spread with the Ymyyakhtakh culture

[Pribislav]

Unrelated:

Has anyone manually checked the Y-calls of I1961? He is that previously published twist sequence sample we now have context for, originally claimed to be N-L1026. However, I saw someone report that he is actually N-Y9023; a subclade also common in Uralic speakers. This result would make a lot more sense as N-L1026 (N-Z1979 as a whole) are intimately associated with Yakutia_LNBA admixture and initially formed/spread with the Ymyyakhtakh culture

I1961; 4239-4002 BC; Sosnovy-Mys, Irkutsk Oblast, Russia; Serovo-Isakovo_LN; N-TAT>F1419>L708>Z4863>Y9023* (xY9022)

This is the first Y9022 sample in aDNA record, and belongs to a very early split, with only 6 derived and 44 ancestral SNPs at Y9022 level. I merged all Y9022 SNPs from both FTDNA Block Tree and YFull for a total of 55 SNPs, 50 of which were covered in his four BAMs. Considering formation and TMRCA dates, the MRCA of I1961 and the ancestor of all living Y9022 people lived ~400 years after the formation date, which would make 5500-5200 BC the earliest date I1961 could've lived. Autosomal analysis by Kale suggests he could've lived more than 1000 years later (post-4500 BC). Six derived SNPs at Y9022 level are: Y9023, Y9281, Y10893, Y148389, Z35164 and BY10351.

Here's the spreadsheet with all relevant calls: https://docs.google.com/spreadsheets/d/1...ue&sd=true

[Jaska]

Very informative article, and fortunately cautious concerning language:

"A large team of Uralicists recently proposed that early Uralic populations were involved in the ST phenomenon, which catalyzed a rapid westward expansion ~4 kya along river networks. While our results are consistent with this scenario, they cannot more precisely inform the question of the location of the Uralic homeland, as they are compatible with multiple hypotheses."

On linguistic basis, it seems impossible that the earliest stage of Seima-Turbino Network in the east could have anything to do with Uralic languages, but for the European part it seems possible. More about this in my forthcoming article "On locating Proto-Uralic" in Finnisch-Ugrische Forschungen 68 (2023).

"Yakutia_LNBA ancestry is present in virtually all known present and ancient Uralic speakers, accounting for all of their East Asian ancestry and decreases in proportion from East to West. Geographically proximal neighbors of Uralic-speaking populations have less or no Yakutia_LNBA ancestry or have other East Asian ancestries in addition to Yakutia_LNBA ancestry. - - Samples from the Late Bronze Age, Iron Age, and Medieval period of the Yenisei River Basin, the Altai region, Western Siberia, and Eastern Europe would be critical in connecting more ancient populations with historically attested Uralic-speaking groups. Such sampling would paint a clearer picture of the dispersals of later, admixed populations carrying Yakutia_LNBA ancestry, and might provide an opportunity to trace such movements to their origin."

Interesting results. Hopefully more samples can be analyzed from relevant regions.

[PopGenist82]

@ Jaska

What is there from perceived loan word strata which can alter the Siberian origins of speakers of Uralic languages ?

[Jaska]

What is there from perceived loan word strata which can alter the Siberian origins of speakers of Uralic languages ?

The problem is that there is zero positive evidence for LATE Proto-Uralic in Siberia. All the proposed evidence for that homeland concerns several millennia earlier stages (Ural-Altaic, Ural-Eskimo etc. connections), and moreover, they contradict with west-pulling evidence (Indo-Uralic connection [connection being here either relatedness or contact]). Therefore, we do not know whether such ancient evidence is even real or due to chance. In any case, so early processes are equally irrelevant for Late Proto-Uralic as the location of LPU is for Proto-Finnic (in Estonia) or Proto-Samoyedic (in South Siberia).

The loanword evidence (especially early Indo-Iranian reconstruction stages) and paleolinguistic evidence allow no LPU homeland east of the Urals, and they require Samoyedic being for a long time adjacent to other Uralic branches. But we can go through the evidence after my article is published.

[PopGenist82]

What is there from perceived loan word strata which can alter the Siberian origins of speakers of Uralic languages ?

The problem is that there is zero positive evidence for LATE Proto-Uralic in Siberia. All the proposed evidence for that homeland concerns several millennia earlier stages (Ural-Altaic, Ural-Eskimo etc. connections), and moreover, they contradict with west-pulling evidence (Indo-Uralic connection [connection being here either relatedness or contact]). Therefore, we do not know whether such ancient evidence is even real or due to chance. In any case, so early processes are equally irrelevant for Late Proto-Uralic as the location of LPU is for Proto-Finnic (in Estonia) or Proto-Samoyedic (in South Siberia).

The loanword evidence (especially early Indo-Iranian reconstruction stages) and paleolinguistic evidence allow no LPU homeland east of the Urals, and they require Samoyedic being for a long time adjacent to other Uralic branches. But we can go through the evidence after my article is published.

Linguistic evidence is in the eye of the beholder . 
Anyway, Germanic loans have been overrated. The “early IE loans” can come from anywhere, and clearly the indo iranian ones come from east of urals. The upper Siberian scenario is any case consistent with linguistic data, appropriately explains the position of Samoyedic as well as the Yukaghir loans. 
The indo-Uralic connections must have another explanation than genetic descent.  

More importantly, a western homeland is simply non tenable from an anthropological perspective because there’s no transmission chain. . And that’s what matters- these were real people who shared enough genetic exchange and culture in order to share their language

To be honest, this S-T is too modest, it doesn’t go far enough to appropriately highlight the origin of early FU speakers

[Andar]

What is there from perceived loan word strata which can alter the Siberian origins of speakers of Uralic languages ?

The problem is that there is zero positive evidence for LATE Proto-Uralic in Siberia. All the proposed evidence for that homeland concerns several millennia earlier stages (Ural-Altaic, Ural-Eskimo etc. connections), and moreover, they contradict with west-pulling evidence (Indo-Uralic connection [connection being here either relatedness or contact]). Therefore, we do not know whether such ancient evidence is even real or due to chance. In any case, so early processes are equally irrelevant for Late Proto-Uralic as the location of LPU is for Proto-Finnic (in Estonia) or Proto-Samoyedic (in South Siberia).

The loanword evidence (especially early Indo-Iranian reconstruction stages) and paleolinguistic evidence allow no LPU homeland east of the Urals, and they require Samoyedic being for a long time adjacent to other Uralic branches. But we can go through the evidence after my article is published.

These early IE loanwords are really not good evidence for Proto-Uralic from west of Urals. We have now evidence for early Indo-Iranian pioneer groups deep in Siberia and North Russia (east of Urals Satyga) around 2000 B.C and in 3000 B.C we had Afanasievo-related groups in Xinjiang and surrounding regions.

Also based on that new papers the connection between Uralics and Seima-Turbino was fairly weak and it definitely was not a phenomenon started by Uralics. But seems that on the northern and very northeastern periphery there were was a contact zone with Uralic-related people and these contacts might have started some dynamic which helped Uralics expanding westwards. But way or another the migration seem to have happened via a quite northern route and not much WSHG or Steppe_MLBA was picked up despite it dominating Seima-Turbino and even showing up in Satyga in mixed Steppe_MLBA+WSHG+Yakutia_LNBA people. Permics show high Steppe_MLBA and Ugrics WSHG afaik but this seems to be from much later LBA/IA admix with locals and steppe groups.

[Jaska]

PopGenist82:
“Linguistic evidence is in the eye of the beholder.”

This claim always comes from people who know nothing about the methods of historical linguistics. There is no method in genetics or archaeology to study language and locate linguistic homeland, as we all know. All those disciplines can do is to try to find matches for linguistic results (if they want to concern language at all).

PopGenist82:
“Anyway, Germanic loans have been overrated. The “early IE loans” can come from anywhere, and clearly the indo iranian ones come from east of urals.”

How are they overrated? True, archaic Indo-European loanwords could come from several branches. Indo-Iranian developed in Europe, and it only spread beyond Southern Trans-Urals ca. 2000 BCE. This is a universally accepted result (excluding some True Believers in India).

PopGenist82:
“The upper Siberian scenario is any case consistent with linguistic data, appropriately explains the position of Samoyedic as well as the Yukaghir loans.“

I just explained to you, how important it is to separate evidence concerning different moments of times. There is no evidence whatsoever supporting LATE Proto-Uralic homeland in Siberia, although very distant Pre-Proto-Uralic might be located there. Samoyedic alone can prove nothing. Uralic loanwords in Yukaghir need not be earlier than post-LPU, but again possible distant relatedness could precede LPU by several millennia (and therefore being irrelevant for locating LPU).

PopGenist82:
“The indo-Uralic connections must have another explanation than genetic descent. “

No matter whether it is genealogical relatedness or contacts, the evidence for location is the same: go west. And as long as there is no reliable method to separate true from false concerning these competing distant connections, nothing certain can be said.

PopGenist82:
“More importantly, a western homeland is simply non tenable from an anthropological perspective because there’s no transmission chain. . And that’s what matters- these were real people who shared enough genetic exchange and culture in order to share their language.”

Good luck for trying to prove your method reliable. Real geneticists understand that language cannot be seen from DNA.

PopGenist82:
“To be honest, this S-T is too modest, it doesn’t go far enough to appropriately highlight the origin of early FU speakers.”

This sounds interesting. Could you write open what do you mean?

[Jaska]

Andar:
“These early IE loanwords are really not good evidence for Proto-Uralic from west of Urals. We have now evidence for early Indo-Iranian pioneer groups deep in Siberia and North Russia (east of Urals Satyga) around 2000 B.C and in 3000 B.C we had Afanasievo-related groups in Xinjiang and surrounding regions.”

This is all true.

Andar:
“Also based on that new papers the connection between Uralics and Seima-Turbino was fairly weak and it definitely was not a phenomenon started by Uralics.”

This comment agrees well with the linguistic results.

Andar:
“But seems that on the northern and very northeastern periphery there were was a contact zone with Uralic-related people and these contacts might have started some dynamic which helped Uralics expanding westwards. But way or another the migration seem to have happened via a quite northern route and not much WSHG or Steppe_MLBA was picked up despite it dominating Seima-Turbino and even showing up in Satyga in mixed Steppe_MLBA+WSHG+Yakutia_LNBA people. Permics show high Steppe_MLBA and Ugrics WSHG afaik but this seems to be from much later LBA/IA admix with locals and steppe groups.”

If you try to read language from DNA, you should be very careful. We do not even know the genetic composition of the Late Proto-Uralic speakers. And you know, the only way to ever know that is to find ancient DNA happening to be in the right place at the right time: in the coordinates given by the linguistic results. The other way round it is just unscientific guessing.

Like, if you first see that the Siberian ancestry in the Uralic populations can be derived from Yakutia_LNBA, and then you decide that the Uralic language must have been inherited just from those people with highest frequency of this particular ancestry. Because you could repeat the same procedure for other ancestries seen in the Uralic populations, and every time you would get different ancestry being responsible for spreading the Uralic language. This shows that this method is utterly unreliable.

[Andar]

Andar:
“These early IE loanwords are really not good evidence for Proto-Uralic from west of Urals. We have now evidence for early Indo-Iranian pioneer groups deep in Siberia and North Russia (east of Urals Satyga) around 2000 B.C and in 3000 B.C we had Afanasievo-related groups in Xinjiang and surrounding regions.”

This is all true.

Andar:
“Also based on that new papers the connection between Uralics and Seima-Turbino was fairly weak and it definitely was not a phenomenon started by Uralics.”

This comment agrees well with the linguistic results.

Andar:
“But seems that on the northern and very northeastern periphery there were was a contact zone with Uralic-related people and these contacts might have started some dynamic which helped Uralics expanding westwards. But way or another the migration seem to have happened via a quite northern route and not much WSHG or Steppe_MLBA was picked up despite it dominating Seima-Turbino and even showing up in Satyga in mixed Steppe_MLBA+WSHG+Yakutia_LNBA people. Permics show high Steppe_MLBA and Ugrics WSHG afaik but this seems to be from much later LBA/IA admix with locals and steppe groups.”

If you try to read language from DNA, you should be very careful. We do not even know the genetic composition of the Late Proto-Uralic speakers. And you know, the only way to ever know that is to find ancient DNA happening to be in the right place at the right time: in the coordinates given by the linguistic results. The other way round it is just unscientific guessing.

Like, if you first see that the Siberian ancestry in the Uralic populations can be derived from Yakutia_LNBA, and then you decide that the Uralic language must have been inherited just from those people with highest frequency of this particular ancestry. Because you could repeat the same procedure for other ancestries seen in the Uralic populations, and every time you would get different ancestry being responsible for spreading the Uralic language. This shows that this method is utterly unreliable.

It is not jut the frequency but that this component and associated uniparental markers are the only shared ancestry between all modern+ancient Uralics, which cant be easily linked with a non-Uralic ethno-linguistic group like CWC/Steppe_MLBA. Sure some specific migrations and groups might have spread with more complex ancestry but that would be Post-Proto-Uralic and doesnt make it Proto-Uralic. Just like all Europeans have today EEF but that doesnt make EEF Proto-IE . We dont know the language of WSHG/ANE-Rich Siberians groups and they likely spoke a bunch of early diverged and unrelated languages from West Siberia to Altai but i find it very hard to link them with Uralics when neither Y-DNA nor autosomal DNA matches and you only see (still minority component) WSHG admix in Udmurts and Ugrics i think. These groups also were mainly pastoralists and i dont think this matches reconstructed vocabulary of Proto-Uralic either.