Paleolithic DNA from the Caucasus

[bicicleur 2]

What he says is that " ‘Mbuti’ ancestry represents the total of ‘Deep’ ancestry from lineages that split prior to the366 split of Ust’Ishim, Tianyuan, and West Eurasians and can include both ‘Basal Eurasian’ and367 other (e.g., Sub-Saharan African) ancestry. "

So, there could be "extra" Basal in there, a lineage that split before Basal, as well as any modern SSA or East African in people. Remember the SSA that used to show up in ancient samples? If the algorithm spotted something really old it threw it in SSA perhaps.
He has no sample for these deep lineages that include not only Basal Eurasian but other lineages that split off before Basal. There's also the language about bottlenecked and non-bottlenecked lineages. Which one was which, the Basal or the pre-Basal? What was the bottleneck, what caused it, where?
Looking at this graph I'm struck by the Dzudzuana in the more "eastern" hunter-gatherers, like Ukraine, Norway, Sweden, Russian Mesolithic, Motala and Iron Gates. Movement westward from the Caucasus? Or straight from Anatolia?
Part of that is Basal Eurasian.
You can also see it in the graph of the amount of Basal Eurasian. According to that, the hunter gatherers had what looks like from 1 to 10%, which is just about what the Feldman et al (Krause) paper on Anatolian hunter-gatherer transition to farmer said would be in the ball park for the amount of Basal Eurasian that would need to be in Iron Gates if there had been movement from Anatolia to Europe at that period. Of course, they only found something like 1.6.

well, at least the label 'Mbuti' is unfortunate then, as Mbuti is a pygmy tribe in Congo

the Dzudzuana in eastern HG is always accompanied by AG3, it seems to me at some point some Dzudzuana admixed with some EHG
what is EHG, isn't it WHG + ANE ?

I wonder in how far common west eurasian = villabruna = WHG ? It is not 100 % correct because there should be some drift in between.
This would mean Dzudzuana = 72 % WHG + 28 % BE
But if you use this equasion then the Feldman AHG = 50 % WHG + 50 % Levant Neo would convert into
AHG = 56 % WHG + 39 % Dzudzuana + 5 % Anciant North African
It's a long shot, and it is approximate but at first sight, it makes some sense to me, it's another way to look at it.
I would expect AHG to have more Dzudzuana than Levant Neo, though a combination of both is possible too.

 

[bicicleur 2]

F is in fact ancestor of GHIJK. https://www.yfull.com/tree/CF/
As for G route, there is indeed a huge gap between the time of G formation and its TMRCA. GHIJK first split could have happened around the area you mentioned, or close to it, indeed. In the case of G, ok, it could have been a route from east to west, till the first split of G itself around Armenia. HIJK from west to east doesn't seem to work, since the clades involved don't present such gap, as you suggested. No time to HIJK "jump" from the area around Caucasus to South Asia.
Anyway, I guess I got your hypothesis. Correct me if I'm wrong. You speculate people related to an older wave from Africa, with a now extinct Y-DNA lineage(?), would have lived in South Asia and would have been the original Basal Eurasians. Then GHIJK - related to another wave - came without this component. Already separated, G and H would have gone south and mixed with this older pop, whereas IJK took another route. Later, G and H would have migrated west bringing BE. Is that right?

yes, it's speculation of course, but that is what I meant

 

[bicicleur 2]

I 100% agree with bicicleur about Y-DNA dispersals, but I think the BE in India is rather unlikely. East Eurasian components from India and surroundings show the typically Eurasian inflation of archaic human ancestry. There's also the similiarity of BE to the newly detected Ancestral North African component, as well as the peaks of BE ancestry in modern North Africans.
I find it more likely that early humans who migrated from South Asia to the West would have picked up BE ancestry in the Arabian peninsula or vicinity when they crossed the Persian Gulf. This would also explain the secondary peak of BE ancestry in ancient Western Iran, whence it migrated up the Zagros range into the Caucasus. BE and Ancestral North African likely existed in a continuum in North & East Africa as well as the Arabian peninsula, the Sinai etc. . Earlier human migrations mopped up archaic human populations so BE/ANA only acquired limited & indirect admixture with them.
It would be interesting to test whether haplogroup D like its sister clade is also at least to an extent associated with a reduction of archaic human ancestry, because its present distribution seems to be the result of a migration that went the opposite direction of most early Eurasian population movements that generally went from the East to the West.
The precise nature of these movements will be very difficult to figure out without DNA from India and South-East Asia in any case, as these places should be the primary homelands of Eurasian humans.

the thing is, IJK seems not to have had BE, so if G+H picked up BE, it must have been after the GHIJK split up
my guess is that IJK moved north, maybe upstream the Indus river, while F+G+H moved along the Indian westcoast, and some of them moved even further inland via the Narmada river

 

[Angela]

well, at least the label 'Mbuti' is unfortunate then, as Mbuti is a pygmy tribe in Congo

the Dzudzuana in eastern HG is always accompanied by AG3, it seems to me at some point some Dzudzuana admixed with some EHG
what is EHG, isn't it WHG + ANE ?

I wonder in how far common west eurasian = villabruna = WHG ? It is not 100 % correct because there should be some drift in between.
This would mean Dzudzuana = 72 % WHG + 28 % BE
But if you use this equasion then the Feldman AHG = 50 % WHG + 50 % Levant Neo would convert into
AHG = 56 % WHG + 39 % Dzudzuana + 5 % Anciant North African
It's a long shot, and it is approximate but at first sight, it makes some sense to me, it's another way to look at it.
I would expect AHG to have more Dzudzuana than Levant Neo, though a combination of both is possible too.

I think he probably used Mbuti to get as far back into the tree as possible? I don't know.

Couldn't the AG3 have come either before or after on its own?

 

[Angela]

Basal means that we see ancestral rather than derived SNP's. So, they split off very early, after which the rest shared mutations and drift. These changes in the rest will be shared among them but not with Basal. As Mbuti is the farthest from anybody else - It split the earliest - chances that it shares something *ancestral* - i.e. not mutated or otherwise changed - with Basal rather than with later Eurasian is greater. Hence the possibility to use Mbuti as proxy. However, if the affinity was due to later SSA admixture, than the part that admixted must share *other* drift and mutations with SSA. In other words, in that case it is not so much the ancestral that it shares but the derived. That should be clearly picked up with formal stats, but it doesn't.

From the SI: "It is clear that Sub Saharan African populations lack shared genetic drift between North African and West Eurasian populations, usually interpreted as the Out-of-Africa bottleneck."



Exactly. Also, we should take a look at the Emiran. It is considered ancestral to Ahmarian in the Levant as well as later UP cultures in North-Africa.



Assuming that you refer to Extended Dat Fig. 7: I am quite curious which HG's that are. The only identified ones have 0% Basal.

It's taking me a while to get through the 76 page supplement, in the midst of, you know, life. I'll let you know.

 

[Angela]

I think he probably used Mbuti to get as far back into the tree as possible? I don't know.

Couldn't the AG3 have come either before or after on its own?

Of course, by the time that CHG formed, i.e. 14-15,000 ya, the ANE had arrived in the Caucasus.

Just found this in the Caucasus, which might be pertinent to our discussions:

"If Vestonice16 did indeed have ancestry from a Villabruna-related population then this type of ancestrywas already in both Europe and the Caucasus by ~30-27 thousand years ago, and mixed (in Europe) withthe earliest inhabitants represented by Sunghir3 and in the Caucasus (see below) with Basal Eurasians.The fact that it was also present in western Anatolia at the time of the Neolithic, as well as in easternEurope by the time of the Eastern European hunter-gatherers from Karelia (see below) suggest adistribution of this type of ancestry around the Black Sea from which it could have propagated."

 

[Angela]

Also interesting. I didn't know this at all.

"Finally, we model Russia_Baikal_EN as a 2-way mixture of Han and 15.7±1.6% MA1-related ancestry.This set of ~7-8 thousand year old samples from Lokomotiv contrast with the ~18 thousand year old AG3sample which as we saw above could be modeled as a clade with MA1. It appears that populations of EastAsian-related ancestry appeared in the region in the intervening period14. As we will see below, mixedAG3/East Asian-related ancestry reached West Eurasia as well."

Did they domesticate the horse? Did they bring that as well as ceramics to the party?

 

[Angela]

I think he probably used Mbuti to get as far back into the tree as possible? I don't know.

Couldn't the AG3 have come either before or after on its own?

I guess I was right.

From the supplement:
"Both Dzudzuana and Taforalt are modeled as 2-way mixtures of Mbuti and Villabruna. Mbuti occupies asymmetrical phylogenetic position to all Eurasians populations—splitting off before the differentiation ofwestern Eurasians, eastern non-Africans, and Ust’Ishim6,12,19 from each other—and can thus be used toquantify such ancestry12."

 

[Angela]

Other interesting bits we haven't discussed.

This answers my question as to which of these ancient lineages is bottlenecked, and it is Basal Eurasian. So perhaps the other deep lineage in Taforalt remained in North Africa?

"Whatever the non-African related population that admixed to form Ibero-Maurusians, we can say that it ismost closely related to Dzudzuana and Villabruna, and that with Villabruna as a baseline, Dzudzuanaalready had some deep ancestry and Ibero-Maurusians from Taforalt even more. The admixture graph40model suggests that this deep ancestry was distinct in Taforalt and Dzudzuana, with Taforalt possessingancestry from both an early and a later split, while Dzudzuana possessing ancestry only from the latersplit (the later split corresponds to the original concept of “Basal Eurasians”ref indicating that it largelyderives from the same bottleneck that affected other non-Africans)."

As to whether Villabruna is "pure":

"Villabruna, is also shown as a 3-way mixture in the model of Table S3.3, tracing about half its ancestryfrom Dzudzuana, and the remainder from Vestonic16 and MA1. This is not a priori implausible as allthese sources are earlier than Villabruna. The admixture graph model presents a simpler model forVillabruna as a simple clade, and an unadmixed Villabruna acts as a plausible source for several other We are thus cautious about accepting this qpAdm result at face value aswell. Earlier sampling may reveal whether Villabruna-cluster6 populations existed earlier than ~15thousand years ago."

Fyi, I am very impressed reading about the modeling. On top of everything else, I like that a lot of it is automated, to remove as much subjectivity as possible.

To continue:
"From our analysis of Supplementary Information section 3, we showed that these sources are indeedcomplex, and only one of these (WHG, represented by Villabruna) appears to be a contributor to allthe remaining sources. This should not be understood as showing that hunter-gatherers from mainlandEurope migrated to the rest of West Eurasia, but rather that the fairly homogeneous post-15kyapopulation of mainland Europe labeled WHG appear to represent a deep strain of ancestry that seemsto have contributed to West Eurasians from the Gravettian era down to the Neolithic period."

 

[Angela]

@Epoch,

"It has been suggested that there is an Anatolia Neolithic-related affinity in hunter-gatherers from the Iron Gates14. Our analysis confirms this by showing that this population has Dzudzuana-related ancestry as do many hunter-gatherer populations from southeastern Europe, eastern Europe and Scandinavia. These populations cannot be modeled as a simple mixture of Villabruna and AG3 but require extra Dzudzuana-related ancestry even in the conservative estimates, with a positive admixture proportion inferred for several more in the speculative ones. Thus, the distinction between European hunter-gatherers and Near Eastern populations may have been gradual in pre-Neolithic times; samples from the Aegean (intermediate between those from the Balkans and Anatolia) may reveal how gradual the transition between Dzudzuana-like Neolithic Anatolians and mostly Villabruna-like hunter-gatherers was in southeastern Europe."

He sounds very sure about this. What it would mean is that they don't have "extra" percentages of these ancient lineages, only what was in Dzudzuana. Something like the way that WHG in Southern Europeans is "hidden" in their Anatolia Neolithic ancestry, and what shows as "WHG" is only the "extra" ancestry?

So, when he says the following, he means "extra" WHG, on top of the related ancestry in Dzudzuana?

"Villabruna: This type of ancestry differentiates between present-day Europeans and non-Europeans within West Eurasia, attaining a maximum of ~20% in the Baltic in accordance with previous observations1 and with the finding of a later persistence of significant hunter-gatherer ancestry in the region14,23,24. Its proportion drops to ~0% throughout the Near East. Interestingly, a hint of such ancestry is also inferred in all North African populations west of Libya in the speculative proportions, consistent with an archaeogenetic inference of gene flow from Iberia to North Africa during the Late Neolithic25."

I've read every word of the Supplement and that's all I could find.

More on these ancient lineages:

"The fact that the genetic drift before and after the Basal Eurasian split is estimated similarly by the admixture graph model of Fig. 2 (which uses no archaic samples or Chimp) and Extended Data Fig. 8 68 (which uses archaic ancestry estimated using Altai, Chimp, and Denisova as outgroups) provides two independent lines of evidence for our estimates of these quantities, suggesting that ~2/3 of the drift since the split from East Africans is shared by Basal Eurasians and an additional ~1/3 is shared by non-Basal Eurasian non-Africans. This suggests that the Basal Eurasians (so named because they occupy a basal position in the phylogeny of Eurasians10) did in fact experience most of the common bottleneck shared by Eurasians. (Note also, that if we used the lower (1.6%) estimate of absolute Neandertal ancestry in Ust’Ishim from the f4-ratio, this would imply even more shared genetic drift between Basal Eurasians and other non-Africans, since then f4(Deep, Tianyuan; Ust’Ishim, Chimp)=-0.016*0.436 ≈-0.007.)"

"The other “Deep” lineage found in Taforalt (Fig. 2) experienced only 0.008 units of genetic drift with non-Africans (Fig. 2) and could be plausibly interpreted as having deep presence in (North) Africa. Note that Taforalt and the Neolithic of the Maghreb are well below the regression line (Extended Data Fig. 8) and thus lack more genetic drift with Ust’Ishim than is predicted by their level of archaic ancestry; this is expected if they trace their ancestry from a lineage that is even more deeply diverged than the Basal Eurasians."

 

[Angela]

For those who haven't yet read it, I think the archaeology section of the Supplement has some pertinent information:

"Conclusions:

It seems that the beginnings of the UP in western Georgia are relatively late, compared withthe earliest UP in the Near East and southeastern Europe and they appear to be alreadydominated by the production of bladelets (e.g., Dzudzuana Unit D). Another importantobservation is that while the Caucasus served as a geographic barrier between two MPNeanderthal populations (the Mousterian of the southern flanks which closely resembles theMousterian of the Taurus and the Zagro, and the Late Mousterian of the northern Caucasus,similar to the northern European Micoquian Mousterian (Meignen & Tushabramishvili2006), the early UP assemblages on both sides of the Caucasus Mountains demonstratesimilarities, indicating the dispersal of modern humans throughout the whole region.The proceeding cultural traditions (e.g., Dzudzuana Unit C) do not follow the UP sequence ofwestern Europe or the Near East as previously claimed. In particular, the ‘carinated core’industries found all over the Caucasus region lack any evidence for the presence of the westEuropean ‘classical’ Aurignacian (and see Belfer-Cohen & Grosman 2007; Goring-Morris &Belfer-Cohen 2006, vis à vis carinated artefacts).

"The presence of carinated cores in some sites may indicate a general contemporaneity amongsites in western Georgia, as with the site of Gubs (Amirkhanov 1986) located on the northernslopes of the Caucasus. Yet, there are no typical Aurignacian tool types either among thelithics or among the bone artefacts. The bone and antler implements in Dzu Unit C (as inother UP assemblages in the Caucasus) do not comprise artefacts such as the split-base point,the hall-mark of the west European early Aurignacian. Bone awls, needles, points and the likewere recovered from UP contexts all over the Old World and the same is true for the rarebone beads and decorations. The same is true also for the other sites in Georgia dating to theearly UP (and see Moncel 2013; Pleardeau et al. 2016) as well as sites in the neighboringArmenia and northern Caucasus (and see above).All in all the UP in the Caucasus retains its own local characteristics, differing both fromEurope (no Aurignacian industry) and the Levant (no el Wad points, yet with rich boneartefacts industries) (and see Golovanova et al. 2014)."

 

[holderlin]

This is awesome

Do some of these models contradict each other? I've drunk too much wine to give an example at the moment, but that was my initial impression.

 

[holderlin]

Baikal Early Neolithic ..

Interesting. It coincides with first pottery in the eastern Pontic steppe .. 9 ka.

All that R1a in the neolithic Baikal samples too. Lines up.

 

[holderlin]

looking at the distribution of F, H and K, it's my guess that FGHIJK split in India or the Indus Valley
these splits happened in a very short time, between 48.8 and 45.4 ka
along the Narmada river in India, some 48 ka blade tools from cilindric core stones were found, maybe the oldest in the world

TMRCA of G is 26 ka, same time as Dzdzuana, maybe a coincidence?
where was G between 48 and 26 ka?

in the mean time, I see that CHG and Iran Neo have even more Basal Eurasian than Dzudzuana, so more Basal Eurasian was coming to this area, parallel with Dzudzuana
Basal Eurasian may have been in India 70 ka or more subsequent to the people in Jebel Faya, 129 ka

I agree with this, but mine was just on a hunch based on the Farmers paper.

 

[Angela]

This is awesome

Do some of these models contradict each other? I've drunk too much wine to give an example at the moment, but that was my initial impression.

Read the supplement when you're not drunk.

 

[Maciamo]

I am commenting a bit late on this new paper due to lack of time. If I understood well, this Dzudzuana sample is ancestral to the population of Eastern European Hunter-Gatherers (EHG). Testing of Mesolithic samples showed that EHG were a mixture of Y-haplogroups I2, R1a and R1b, but dominated by haplogroup R1b - notably in the Iron Gates and Latvia as well as the Villabruna sample from Epipaleolithic Italy.

This means that Dzudzuana could belong to Y-haplogroup R1* (as R1a and R1b had not yet appeared 26,000 ybp according to current age estimates.) It's a shame that the pre-print doesn't mention it.

It has been suggested that there is an Anatolia Neolithic-related affinity in hunter-gatherers from the Iron Gates. Our analysis confirms this by showing that this population has Dzudzuana-related ancestry as do many hunter-gatherer populations from southeastern Europe, eastern Europe and Scandinavia. These populations cannot be modeled as a simple mixture of Villabruna and AG3 but require extra Dzudzuana-related ancestry even in the conservative estimates, with a positive admixture proportion inferred for several more in the speculative ones.

So, in other words, when the authors say that Dzudzuana can be modelled as a mixture of Villabruna (R1b from Epipaleolithic NE Italy) and AG3 (14,500 year-old Afontova Gora 3 from Lake Baikal in southern Siberia, which belonged to Y-haplogroup R1b), it is not very different from saying that it is EHG or SHG and therefore related to Y-haplogroup R1, and particularly R1b.

The fact that Dzudzuana is 'more closely related to to early agriculturalists from western Anatolia ~8 thousand years ago than to the hunter-gatherers of the Caucasus from the same region of western Georgia of ~13-10 thousand years ago' can easily be explained by population blending between R1 EHG from Southeast Europe (or even western or northern Anatolia) and those early Anatolian farmers.

 

[bicicleur 2]

All that R1a in the neolithic Baikal samples too. Lines up.

there was also N, C2 and Q
the R1a must have been the R1a-YP1272, which spread the comb-ceramic culture

https://www.yfull.com/tree/R-M459/

 

[bicicleur 2]

Many of the old theories that were floating around receive a lot of support in this paper.

* North Africa as a source of massive populations movements, possible at several points during the Paleolithic & Mesolithic. These movements likely brought haplogroup E (back) to Eurasia.

* North Africa as a significant source of admixture in modern West Africans.

* ANE as a two-way mixture between a West Eurasian- and an East Eurasian source (here represented by Tianyuan). The East Eurasian population likely brought with it haplogroups R & Q from South-East Asia.

the 73 % Dzudzuana in Natufian must come from the Kebaran
https://en.wikipedia.org/wiki/Kebaran

The appearance of the Kebarian culture, of microlithic type implies a significant rupture in the cultural continuity of Levantine Upper Paleolithic. The Kebaran culture, with its use of microliths, is associated with the use of the bow and arrow and the domestication of the dog.[1] The Kebaran is also characterised by the earliest collecting of wild cereals, known due to the uncovering of grain grinding tools. It was the first step towards the Neolithic Revolution.

Taforalt also has 55 % Dzudzuana

E-M35 is probably of African origin, but before it spread 24 ka, it was heavily admixed with the Eurasian Dzudzuna-like DNA

 

[halfalp]

Btw, i found interesting the new Baikal Early Neoilithic component wich is mostly 50% Han and 50% Mal'ta Related. It is found now in prehistoric eastern europe. Doesn't it confirm the old archeological hypothesis that from Northern China / Mandchuria, Comb Ceramic ( and the first Ceramic in general ) and maybe also Millet, roam until eastern europe and finland? It was maybe progressive, but it certainly was both cultural and demic.

 

[halfalp]

I am commenting a bit late on this new paper due to lack of time. If I understood well, this Dzudzuana sample is ancestral to the population of Eastern European Hunter-Gatherers (EHG). Testing of Mesolithic samples showed that EHG were a mixture of Y-haplogroups I2, R1a and R1b, but dominated by haplogroup R1b - notably in the Iron Gates and Latvia as well as the Villabruna sample from Epipaleolithic Italy.

This means that Dzudzuana could belong to Y-haplogroup R1* (as R1a and R1b had not yet appeared 26,000 ybp according to current age estimates.) It's a shame that the pre-print doesn't mention it.



So, in other words, when the authors say that Dzudzuana can be modelled as a mixture of Villabruna (R1b from Epipaleolithic NE Italy) and AG3 (14,500 year-old Afontova Gora 3 from Lake Baikal in southern Siberia, which belonged to Y-haplogroup R1b), it is not very different from saying that it is EHG or SHG and therefore related to Y-haplogroup R1, and particularly R1b.

The fact that Dzudzuana is 'more closely related to to early agriculturalists from western Anatolia ~8 thousand years ago than to the hunter-gatherers of the Caucasus from the same region of western Georgia of ~13-10 thousand years ago' can easily be explained by population blending between R1 EHG from Southeast Europe (or even western or northern Anatolia) and those early Anatolian farmers.

But how can Dzudzuana being ancestral of EHG if Dzudzuana is unrelated with CHG wich is intermediate with EHG and Iran_Neolithic?

So did AG3 belong to y-dna R1b and not mtdna R1b? There is a lot of confusion about this detail from many years now. The wiki state it is mtdna R1b.