search this blog

Friday, September 21, 2018

Dzudzuana Ice Age foragers: a different type of Caucasus hunter-gatherer (Lazaridis et al. 2018 preprint)

Over at bioRxiv at this LINK. Below is the abstract. Emphasis is mine.

The earliest ancient DNA data of modern humans from Europe dates to ~40 thousand years ago, but that from the Caucasus and the Near East to only ~14 thousand years ago, from populations who lived long after the Last Glacial Maximum (LGM) ~26.5-19 thousand years ago. To address this imbalance and to better understand the relationship of Europeans and Near Easterners, we report genome-wide data from two ~26 thousand year old individuals from Dzudzuana Cave in Georgia in the Caucasus from around the beginning of the LGM. Surprisingly, the Dzudzuana population was more closely related 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. Most of the Dzudzuana population's ancestry was deeply related to the post-glacial western European hunter-gatherers of the 'Villabruna cluster', but it also had ancestry from a lineage that had separated from the great majority of non-African populations before they separated from each other, proving that such 'Basal Eurasians' were present in West Eurasia twice as early as previously recorded. We document major population turnover in the Near East after the time of Dzudzuana, showing that the highly differentiated Holocene populations of the region were formed by 'Ancient North Eurasian' admixture into the Caucasus and Iran and North African admixture into the Natufians of the Levant. We finally show that the Dzudzuana population contributed the majority of the ancestry of post-Ice Age people in the Near East, North Africa, and even parts of Europe, thereby becoming the largest single contributor of ancestry of all present-day West Eurasians.

Lazaridis et al., Paleolithic DNA from the Caucasus reveals core of West Eurasian ancestry, bioRxiv, posted September 21, 2018, doi:

See also...

Villabruna cluster =/= Near Eastern migrants


Chad said...

LOL! I guess my tree with Iberomaurusians having no SSA but mixing into SSA was correct.

When_in_Rome said...

Looks like you made some predictions that may be validated:

Regarding this post and your last post:

How related are Villabruna to other Paleo-Europeans?

EastPole said...

"Genes reveal traces of common recent demographic history for most of the Uralic-speaking populations"

epoch said...

Next to this epic paper there is another interesting one out:

We show that most Uralic speakers share a distinct ancestry component of likely Siberian origin, which suggests that the spread of Uralic languages involved at least some demic component.

epoch said...

MtDNA is U6 and N. No Y-DNA mentioned

Samuel Andrews said...

Ancient west Eurasians are seeming more and more similar. This paper argues Dzudzuana-like pops form the foundation of all the ancient Middle Eastern pops from Morocco to Iran.

"Thus, the highly differentiated
populations at the dawn of the Neolithic were primarily descended from Villabruna Cluster
and Dzudzuana-related ancestors, with varying degrees of additional input related to both North Africa and Ancient North/East Eurasia whose proximate sources may be clarified by future sampling of geographically and temporally intermediate populations. "

Dzudzuana is modeled as mostly Villabruna-like. Does that mean all West Eurasians are mostly Villabruna-like? IMO, no way. But, it does show there's shared ancestry in all the Mesolithic/Neolithic lineages that contributed to modern West Eurasians.

Forever, we have all known that when you ignore Basal Eurasian stuff, all the Meso/Neolithic west Eurasian pops traced lots of ancestry back to the same distant Upper Paleolithic populations.

Dzudzuana looks like the populations we have suspected existed. A Basal/Villabruna-like mix with no ANE influence who can be modeled as the main ancestor of all ancient Middle Eastern pops.

Samuel Andrews said...

If you go so far back the "Villabruna/WHG" thing changes meaning. ElMiron & EHG have more relation to WHGs than any of the ancient Middle Easterners who are also modeled as mostly Villbruna/WHG-like. This is not just because of the effect of Basal Eurasian ancestry. It is because the European's WHG ancestry is more related to the Y DNA I-R1b/mtDNA U5b Mesolithic western Europeans.

Samuel Andrews said...

There's a brief mention to mtDNA. No mention to Y DNA.

One had U6, one had 'N.' The U6 is a big deal for obvious reasons. It's a direct maternal link with Epipaleolithic north Africans Taforalt.

Gihanga.Rwanda said...

Chad - The Iberomaurusians are modeled as 45% “Ancient North Africans (ANA)” here. ANA’s closests relatives appear to be Mota and other SSAs.

a said...

What a wonderful opportunity to compare these results with,
Oase1-Neanderthal- Mezmaiskaya Cave Neaderthal and Steppe Neaderthal.
Same format as the previous paper --Genomic insights into the origin of farming in the ancient Near East, would be great.
West Eurasians harbour significantly less Neanderthal ancestry than East Asians19-21, which could be explained if West Eurasians (but not East Asians) have partial ancestry from a source diluting their Neanderthal inheritance22. Supporting this theory, we observe a negative correlation between Basal Eurasian ancestry and the rate of shared alleles with Neanderthals19 (Supplementary Information, section 5; Fig. 2). By extrapolation, we infer that the Basal Eurasian population had lower Neanderthal ancestry than non-Basal Eurasian populations and possibly none (ninety-five percent confidence interval truncated at zero of 0-60%; Fig. 2; Methods). The finding of little if any Neanderthal ancestry in Basal Eurasians could be explained if the Neanderthal admixture into modern humans 50,000-60,000 years ago11 largely occurred after the splitting of the Basal Eurasians from other non-Africans.

So the question- how much Neanderthal did these samples harbour remains to be posted.

Davidski said...


ANA’s closest relatives appear to be Mota and other SSAs.

All of the main human clades appear closely related in that part of the tree because that's where all of the really deep bifurcation happens.

So if you're suggesting that ANA is a Sub-Saharan clade, then nope, it's something all of its own.

Gihanga.Rwanda said...

David - How would you define a Sub-Saharan clade in this context? According to this paper, ANA (a population with a deep presence in North Africa, plausibly) is supposed to lack any Neanderthal ancestry and have diverged sometime after Mota, but early enough to have not experienced much of the genetic drift shared by non-Africans, including Basal Eurasians. So my question to you is what makes Mota “SSA” and not ANA? If the OOA bottleneck is what defines “non-Africans”, ANA clearly doesn’t fit the bill.

Davidski said...


How would you define a Sub-Saharan clade in this context?

Any clade indigenous to south of the Sahara.

So my question to you is what makes Mota “SSA” and not ANA?

The Mota clade can't be defined as ANA because it's a clearly a different, deeply diverged sister clade. Strictly speaking it's an East or Northeast African clade, so it doesn't have to be called Sub-Saharan, but I think it can since it's indigenous to a region deep in Africa.

But these are just semantics. The point is that ANA isn't a Sub-Saharan clade, irrespective of how anyone defines the Mota clade.

If the OOA bottleneck is what defines “non-Africans”, ANA clearly doesn’t fit the bill.

It doesn't because it's a unique North African clade.

Gihanga.Rwanda said...

David - I don’t necessarily disagree with anything you just said. We are discussing semantics. I just wanted to confirm that you weren’t for some reason grouping other African clades together exclusive of ANA on any grounds besides geography. To put it briefly, a pan-African or SSA clade don’t exist; instead what you have are multiple “unique” and “different, deeply divergent clades” that brached off before the OOA bottleneck and have contributed to modern Africans (both above and below the Sahara) to varying degrees... ANA being one them.

Davidski said...


By the way, Dzudzuana, what a great name for a cave and an ancient human population!

Michalis Moriopoulos said...

Pronounced Joo-joo-ah-nah, if my understanding of Georgian phonology is correct.

Samuel Andrews said...

The admixture proportions in Figure 6 for modern/ancient west Eurasians isn't completely correct. It used good outgroups but nonethless it can't be right. It's mostly right if we assume all ancient Mid Eastern pops were mostly Dzudzuana-like.

I'd like to see Yamnaya results. Because, there's no way northern Europeans are only 10% Villrubna but 60-70% Dzudzuana. In this model both Yamnaya & Neolithic Europeans should have a big chunk of Villarubna.

Ric Hern said...

Any guesses which Y-DNA Haplogroup these people were ?

George Okromchedlishvili said...

@Mike the Jedi:
No, ძუძუანა is pronounced as dzu-dzu-ana.
Roughly translates as :"of tits", or just "tits spot"

Davidski said...

@Samuel Andrews

Yeah, the bar graph showing the ancestry proportions is a simplification of what actually went down, and honestly parts of the paper like that one look hyperbole to me, but they do make this qualifying statement at the end...

We caution that the inference of Dzudzuana-related ancestry as the core component of ancient and present-day West Eurasia does not constitute proof for migrations specifically from the Caucasus: given that this is the only ancient DNA data from this time period and broad region, the geographical and temporal extent of this population and its relatives remains unknown. Both in its past (formed by admixture with Basal Eurasians), and in its future (admixing with populations from Africa, Europe, and Siberia in post-glacial, Neolithic, and later periods), Dzudzuana stands in the middle of an ongoing process of admixture of diverse elements from which West Eurasians, in all their diversity, eventually formed.

A said...

Did you see this paper on Sardinian DNA? 18% have R1b (R-M268) but ‘steppe’ ancestry is very low.

A said...

Genomic history of the Sardinian population:

epoch said...

The 9000 year old Chan do Lindeiro sample, published the Iron Gates paper [1] where its difference from other Mesolithic was already noted, comes out as Magdalenian in the ADMIXTURE output. That is really surprising, to say the least.


Shaikorth said...


Do you mean qpAdm output?
That's indeed surprising, as is the Dzudzuana in Sweden_Mesolithic and Sidelkino but not in Karelia_HG. If that holds there's an odd fragmentary pattern of Basal spread in Europe pre-Neolithic.

Angantyr said...


They don't specify what "Sweden_Mesolithic" is, they just refer to Günther 2018, which has four new/resequenced Swedish HGs from Stora Förvar and Stora Bjers.
Three of these, SBj, SF9 and especially SF11, do show hints of contamination, so if they used all individuals and not just the higher quality SF12, there could perhaps be a risk that Dzudzuana in Sweden_Mesolithic comes from contamination.

alobrix said...


Chan do Lindeiro sample is from the mountains of actual Galicia.

If you make a PCA with iberian or european populations from Global25 averages, and then a minimum spanning tree, you see this:

There is a preferential connection between Galicia and GoyetQ116-1.

Are there a special relationship, tiny but perceptible, between the nowadays galicians and the remote aurignacians?

I think we need more formal approaches to test this surprising possibility.

epoch said...


Yes, qpAdm, according to the text. Extended Data Figure 6.


Is that sample - It was originally called Chan_Meso in the publishing paper - available?

Davidski said...


I don't have the genotype data and I won't go out of my way to get it.

Matt said...

Use of their Fst data in Extended Data Table 1 through the PCoA and clustering methods in PAST3:

Unfortunately no real outgroups like Mbuti / Papuan present in data to cross check how position of Dzudzuana relates to lower overall drift, vs phylogenetic centrality (though of course more dubious about whether all African groups are "true" outgroups to Near East following this paper).

Samuel Andrews said...

Some Middle Neolithic, Chalcolithic Iberian farmers have Elmiron ancestry according to G25. Also, their Anatolian ancestors had less IranNeo stuff than Barcin & other Neolithic Europeans.

Open Genomes said...

@David, I take it that the SNPs for Dzudzhuana aren't available yet. Do you have the Anatolian Eneolithic and Neolithic samples in Global 25 yet?

Davidski said...

Yes, Barcin, Tepecik, etc.

Matt said...

Using those Fst PCoA that I posted up yesterday, here are some quick nMonte runs with that data:

Here's the PCoA on Fst data if you want to do anything yourselves:

If allowed to compete with all other populations, Anatolian takes about 25% ancestry from Dzudzuana and 61% from PPNB. Using only Upper Paleolithic sample set, 52% ancestry from Dzudzuana, 30% from Natufian, 14% from CHG. PPNB models as Anatolia+Natufian with no Dzudzuana edge, or removing Anatolian about 50% Natufian, 32% Dzudzuana.

I'd say best guess for me is CHG and Anatolian is about 50% ancestry from Dzudzuana at this point - being wary of tendency to describe Iran_N, Natufian in terms of Dzudzuana+Extra Basal Eurasian ghost population, as the paper seems to do. (Implying about roughly 20-40% of Dzudzuana like ancestry in Europe today?).

It still seems like even this is not a good fit though; Dzudzuana may end up being an ancient population that didn't contribute much, with ghost populations richer and poorer in Basal being what actually contributed, and similarities in apparent Basal Eurasian level to Anatolian mere coincidence. (Almost extinct pops common in Europe, why not Near East?). It'll be nice to repeat this exercise in future with a bigger Fst matrix with more outgroups.

Biggest groundbreaking facet of this paper is dating Basal Eurasian ancestry in Near East, claims about Dzudzuna being the "core" of West Eurasian ancestry do seem overblown, yeah.

Matt said...

Does seem like a mild Gravettian affinity for Dzudzuana in Fst, relative to Anatolian and assuming equal affinity for WHG+Aurignacian net of drift -

Doesn't seem to show up strongly in their formal stats though. I wonder why?

epoch said...


"Doesn't seem to show up strongly in their formal stats though. I wonder why?"

Maybe due to the LGM isolation? DzuDzu and both Gravettian and Aurignacian are pre-LGM, whereas WHG are post-LGM.

pnuadha said...

"ANA’s closests relatives appear to be Mota and other SSAs." - Gihanga

Then you dont know how to read the graph. ANA's closest relatives are Eurasians.

"So if you're suggesting that ANA is a Sub-Saharan clade, then nope, it's something all of its own."

This comment doesnt make a whole lot of sense. If you call ANA its own thing you might as well call basal Eurasian its own thing, East non African its own thing, ANE its own thing, and keep going down the list till you reach the end you happen to choose.

The point is that ANA forms a clade with the rest of the Eurasians. Id classify it as MENA, a subset of Eurasia, since thats where it evolved and mostly resides.

Ric Hern said...

Now I wonder if it is Haplogroup G or I in that Population ? If it is G then is G closer to I than it is to J ? What does this say about the proposed Makro-Haplogroups IJ, IJK and GHIJK ?

Gihanga.Rwanda said...

Colin Welling - Read the supplementary documents. Unlike Basal Eurasians, ANA doesn’t share much of the common drift characteristic of non-Africans relative to Mota and the bulk of the ancestry in Yoruba. It’s a distinct ancestral component likely indigenous to North Africa; we don’t know it’s full geographic extent, but if I had to bet, Paleolithic and early-Neolithic populations from NE Africa along the Nile Valley were probably ANA-rich. It’s obviously intrusive to West Asia, since the Natufians and populations with Natufian-related ancestry are the only non-Africans to carry such admixture. Intentionally or not, you’re being disingenuous.

capra internetensis said...

Hmmm, ANA ghost kinda fits with hypothetical Para-Eurasian ghost (African sister to OOA, associated with hg E), but def not really in this model as contributes too little to Yoruba and not at all to Mota. OTOH the model is surely wrong for SSAs, which are not really the focus I guess - Mbuti is not unadmixed, and where has the Basal African in Yoruba got to? That was supported as real by the more-or-less matching archaic ancestry found via a different method.

Dzudzuana age is consistent with momi2 estimate of BE x WHG-like mix in EEF ancestors occurring ~33 (11-41) kya, but this model has large contribution of BE not much diverged from Crown Eurasian, while the momi2 model had a small contribution of more divergent BE. I don't suppose it would be possible to model BE as upstream of ANA?

pnuadha said...

Gihanga, This dispute started because you said ANA’s closest relatives appear to be Mota and other SSAs

You are factually, not semantically, wrong. ANA is closer to Eurasians than sub Saharan Africans.

ANA formed outside of sub Saharan Africa and belongs to a clade that excludes sub Saharan Africans. There is no genetic or geographic basis to support your grouping of ANA with other sub Saharan Africans.

Unlike Basal Eurasians, ANA doesn’t share much of the common drift characteristic of non-Africans relative to Mota and the bulk of the ancestry in Yoruba

You're just saying that basal Eurasian is further down the tree... so is East Eurasian, doesnt change a thing.

ANA is North African and if you talk about a greater family, it's with Eurasians, not SSAs.

Gihanga.Rwanda said...

Colin Welling - Did you read the supplementary materials, it’s much more informative than the abstract? In this particular model, Basal Eurasian shares .021 (~66%) units of common drift with the main Eurasian branch (West + East Eurasians) exclusive of Mota, while ANA only shares .008 (~25%) units of drift. How exactly do these stats support your position? This pattern holds true if you use the main “Yoruba” ancestry (that 87%) as the starting point to measure drift between Africans and non-Africans. ANA is another example of deep population substructure in Africa independent or inclusive of Eurasian diversity, depending on how you look at it. And as I noted earlier, ancestral African populations do not form tight clades with each other.

mad green said...

ANA -“Deep” ancestry- diverged from other non-Sub-Saharan Africans before the Basal Eurasians did...however this Deep ancestry group, which nevertheless forms a clade with the new Eurasian lineages (Basal and non-Basal), does not clearly seem to have undergone the bottleneck event according to this paper.
so if ANA is anything it's basal to basal Eurasian...this paper was pretty much clear about this part ,so in other words people should stick to the DATA and keep their emotions and any racist agenda they've got away such discussions.

Gihanga.Rwanda said...

Mad Griss and nuadha (are you Colin Welling?) - It’s hard to have a discussion with people who lack reading comprehension skills. All Africans are basal to Basal Eurasian, have you heard of a concept called “Out of Africa”? So ANA forms a clade with Eurasians including Basal Eurasian to the exclusion of other Africans; you widen that some more, you can say pretty much the same thing for Mota, the core ancestry in Yoruba, and so on. Those stats matter speak to ANA’s genetic affinity to Eurasians relative to other Africans. This is simplifying things but in line with the study, but if ANA is ~20% removed from Mota (East African), approximately equidistant from main Yoruba and Basal Eurasian (~35-40%), and ~60% from main Eurasians, what does that say about ANA’s place in the human lineage? ANA is distinct African lineage that contributed to some West Asians, hence why you’re butt hurt. My African vs. non-African definition is pretty simple: lineages that primarily diverged before and after the OOA bottleneck. ANA is “African”, Basal Eurasian is not.

Gihanga.Rwanda said...

We are just scratching the surface with this study; more aDNA from Africa will probably reveal more ancient populations with varying degrees of affinity to OOA populations, some closer like ANA and Mota, while others more basal (e.g. 2000kya SA), and much more basal (e.g. Basal African). Things are about to get a bit complex and I am sorry that we can’t box all Africans into a tight little box of homogeneity.

pnuadha said...

Chad said LOL! I guess my tree with Iberomaurusians having no SSA but mixing into SSA was correct

This is entirely accurate.

You replied ANA’s closests relatives appear to be Mota and other SSAs.

Wrong, wrong, wrong.

ANA is closer to Eurasians than all sub Saharan Africans. There isnt much else to discuss other than you being triggered by anybody saying ANA is not sub Saharan African.

My African vs. non-African definition is pretty simple: lineages that primarily diverged before and after the OOA bottleneck.

Good for you; non African is a perfectly valid geographic designation and genetic clade. Non sub Saharan African is a perfectly valid geographic designation and genetic clade.

To recap, ANA is North African, not SSA. ANA is more related to Eurasians than any SSA group and vice versa. Unless you disagree here, you have no real reason to argue.

In terms of the way I personally think of ANA, I find non sub Saharan African a useful designation for geographic and historical reasons. To me, the fact that it forms a genetic clade with Eurasians is a lesser point than the fact that it is genetically separate from SSA and mostly resides in the Middle East and North Africa for all history.

Mark Moore (Moderator) said...

KingJohn at Anthropogenica wrote " (Epipaleolithic Natufians and early Neolithic Pre-Pottery Neolithic
PPNB farmers) can also be modeled as a mixture of Dzudzuana and Mbuti, with a little more deep ancestry in Natufians (~11%) than in PPNB (~7%) on top of the deep ancestry of Dzudzuana.

Thus,Villabruna→Dzudzuana/Anatolia_N→PPNB→Natufian→Taforalt represents a
cline of increasing deepancestry (and decreasing Villabruna-related ancestry) in what was previously termed the South/West West Eurasian interaction sphere12"

So if I get that right he is saying they can all be described in terms of "Deep" ancestry without reference to what is called "Basal Eurasian" which is less Basal than Deep. Any comments on that folks? Can "Deep" explain Dzudzuana without input of Basal Eurasian? Also is the Basal Eurasian which shows up < 15K ago anything different from "Deep"?

Arch Hades said...

"Too recap, ANA is North African, not SSA. ANA is more related to Eurasians than any SSA group and vice versa. Unless you disagree here, you have no real reason to argue."

I assume you're going by their phylogenetic tree? So in that case Mota who is Sub-Saharan is more similar to ANA and Eurasian than he is to other Sub-Saharans?

Ric Hern said...

Bit of topic.

Any info on Sidelkino441 Mesolithic sample ? Just asking because maybe I missed something...

Karl_K said...

When are we going to get some non-admixed Basal Eurasian genomes? Now that would be very interesting.

Puree said...

I'm wondering whether any Dzudzuana samples ( ever made it to the G25 database? I am read this blog from time to time but not every word so my apologies if I missed the information somewhere. Also I have checked the ancient G25 DB using the term Dzudzuana and did not find so if it is there then my search wasn't working or it may be under a different name. Thanks for the help.

Davidski said...


The paper and genomes haven't been published yet. I don't know when that will happen.