Sunday, May 21, 2017

Do Yamnaya Autosomal Genetics Derive From Caucasian Mail Order Brides?

Davidski at Eurogenes proposes a model for the formation of the Yamnaya genotype that makes a couple of key assumptions set forth below (emphasis in the original): 
- It's likely that low population densities in Eastern Europe during the Eneolithic ensured the rapid spread and rise of admixture from the Caucasus across much of the Pontic-Caspian steppe, which then plateaued at around 50% during the Yamnaya period, when population densities on the steppe may have become high enough so that continued gene flow from the Caucasus no longer had much of an impact. 
- The process that led to the Yamnaya genotype eventually led to its extinction by the Late Bronze Age, due to the large scale spread of Middle Neolithic European farmer ancestry across the entire Pontic-Caspian steppe, probably from its western half, resulting in the formation of the Steppe_MLBA genotype, exemplified by the Sintashta and Srubnaya people. 
- Ancient DNA suggests that Bronze Age steppe groups were highly patrilocal, and if so, it's likely that most of the mixture on the steppe at this time was facilitated via female exogamy (i.e. foreign brides), which would explain the lack of typically Caucasian Y-haplogroups, such as J2, in Bronze Age steppe and derived ancient groups sampled to date, such as the Corded Ware people and eastern Bell Beakers.
He also notes that Steppe people were highly mobile and that the circumstances encouraged admixture of Steppe peoples. There are some key unstated points are key as well (my statements, not a quotation):
* Steppe men are overwhelmingly Y-DNA R1, while this is rare among Caucasian men of any time period. Actually, Yamnaya culture men are predominantly Y-DNA R1b in the case of the Yamnaya, although the ancient DNA of Steppe men further to the North and in the Corded Ware Culture (which also had a more northern orientation) overwhelmingly reveals Y-DNA R1a.  (In both cases, of high derived version of each haplogroup, not the most basal versions.)
* Yamnaya autosomal genetics can be decently approximated by blending ancient DNA from Eastern Hunter-Gatherers (60%) and Caucasian Hunter-Gatherers (40%). This is what makes it necessary to hypothesize an influx of foreign Caucasian brides. See, e.g., here (migration from the Steppe paper) and here (earlier Eurogenes post analyzing ancient DNA samples). Genetic evidence of Caucasian admixture in ancient DNA is absent on the steppe until the Eneolithic (i.e. the Copper Age).
* Early Neolithic Caucasian farmers appear to have been largely derived from Caucasian Hunter-Gatherers, in situ, rather than largely replacing existing hunter-gatherers in places where they expanded as first wave Neolithic people in Europe derived from Anatolian Neolithic people did.  (See also here).
* There is strong physical evidence of patrilocality and large numbers of wives who migrate to their husband's homes from considerable distances from the remains of central European Bell Beaker men who have significant steppe ancestry. Other physical anthropology studies of Bronze Age European women have reached similar conclusions.
There is also anthropological corroboration for this hypothesis (again, my observations and not a quotation):
* Much of the metallurgy technology package that will subsequently be spread across Europe by people with significant steppe ancestry was probably first developed in the Caucasus mountains area. 
* The source of the steppe farming and herding package is less obvious, with David Anthony making a fairly solid case from archaeology in The Horse, The Wheel and Language for a basically Balkan origin of farming and herding technologies on the steppe. In general, there is considerable confusion among historical linguists and pre-historians over why Proto-Indo-European seems to have so many horticulture specific  and maritime shared root words for a society often conceptualized as made up of herders. Hunter-gatherers tend to transition much more easily to a pastoralist lifestyle (i.e. herding) than to horticulture that requires a much greater lifestyle and worldview change. But, the Early European Farmer autosomal DNA does not show up in Steppe populations until the Middle Bronze Age, so it is hard to seem Balkan farmers playing an important demic role in the formation of the steppe genotype. 
* The Caucasus mountains is also one possible source of a religious theme of a fire cult (also here) that is visible in the Zoroastrian religion which was established by Indo-Iranians (although also possibly in BMAC temples in the Harappan trade network and there is some linguistic evidence to support the idea that they shared a language with the Harappans from substrate words in Indo-Iranian and Indo-Aryan languages), although the center of gravity of this cult seems further east and later in time to be a good fit. A possibly residual fire cult can also be found in the cult of Hephaestus in the Aegean (with parallels in both Indo-European and non-Indo-European deities elsewhere in the region), although Hephaestus may have pre-Mycenean Greek roots as this blacksmith god is closely associated with the island of Lemnos which was one of the last non-Indo-European language hold outs in the region which was spoken until the 6th century BCE. The God's deformity is characteristic of Bronze Age rather than Iron Age smiths, who were often poisoned by arsenic associated with tin needed to make Bronze. 
The link between Caucasian brides and the people of the Pontic-Caspian steppe which is pretty much the only credible source of the main expansion (at least) of the Indo-European languages, however, is complicated. 

Did Caucasian brides and their children come to speak their husband's steppe language?  Or, could Indo-European be a Caucasian origin language passed from mother to child on the steppe?

The Caucasian mountains are currently home to several families of non-Indo-European languages that appear in a variety of respects (e.g. elaborate grammar and uncommon phonemes) to relict languages from a very long time ago (probably back at least to the early Neolithic in the region) that have not been heavily influenced by language contact or first language learners in highly isolated mountain valleys. The evidence from ancient mtDNA tends to confirm population genetic continuity in the Caucasus, although the ancient Y-DNA evidence seems to show more change over time.

So, the former hypothesis, that brides and their children came to speak their husband's language, seems more likely and has some corroboration in the pattern observed in similar situations that are historically attested.

One possibility is that Southern steppe people were not originally Indo-Europeans and had a strongly Caucasian influenced language (possibly related to Minoan and Basque), but that northern steppe people were Indo-Europeans. The fact that early Iberian Bell Beaker culture people seem to largely lack steppe ancestry, even though non-Iberian Bell Beaker people have it to a great degree, however, complicates the story.

New ancient DNA to come from ancient Minoans (non-Indo-European pre-Greeks of Crete), Mycenaeans (the first Greek speaking people of the Aegean), Maykop culture people (a technological source culture at the foothills of the Caucasus mountains near the Black Sea coast), and Harappans (pre-Indo-Europeans of the Indus River Valley civilization) will help shed some light on the different possible hypotheses in the near future.


Ryan said...

"One possibility is that Southern steppe people were not originally Indo-Europeans and had a strongly Caucasian influenced language (possibly related to Minoan and Basque), but that northern steppe people were Indo-Europeans. The fact that early Iberian Bell Beaker culture people seem to largely lack steppe ancestry, even though non-Iberian Bell Beaker people have it to a great degree, however, complicates the story."

I think the simpler explanation is that Iberian Bell Beakers moved to central Europe and assimilated Corded Ware Culture locals before having a secondary expansion. That would fit with all of the data. People just assume it's only IE speakers who went around assimilating others, but I see no reason for that to necessarily be the case.

I suspect though that Yamnaya got its Indo-European language from it's ANE heritage, whereas it got its R1b Y-chromosomes from its WHG heritage. I think R1a and Q are the markers that were linked to IE's older ancestors, not R1b, which would have arrived in Europe quite a bit earlier.

Consider that all the ancient samples we have of R1b are only from populations with WHG ancestry. It's the only common factor between them. Many lack any appreciable Steppe ancestry, so I don't think it makes sense to view R1b and steppe ancestry is very closely tied. There's R1b Iberian Bell Beakers who lack steppe ancestry even.

I think the data fits with Bell Beakers initially carrying high frequencies of I2 and R1b spreading to central Europe with a relatively modest autosomal impact, and through a series of founder effects one clade of R1b becomes dominant in the group that reaches Central Europe and mixes with CWC. Then a secondary, more demographically important expansion comes from that mixed group (and IE speakeres influenced by Bell Beakers do their own expansion as the Unetice culture.

The Bell Beakers were almost entirely WHG on their Y-chromosomes though. I think the bigger question is this WHG resurgence in the middle Neolithic that seems to have impacted everywhere from Portugal to the Ukraine. What caused that? Is there some WHG network/culture that the Bell Beakers just managed to make use of even?

andrew said...

The empirical link in both modern and ancient DNA between Y-DNA R1b (the derived version found, e.g., in Western Europe and the Yamnaya, not the more basal R1b-V88), and autosomal steppe ancestry is extremely strong, even if not 100% perfect. And, WHG ancestry is widespread in populations such as first wave Neolithic people and Mesolithic populations, in which the derived Y-DNA R1b (e.g. R1b-M269) is completely absent.

Y-DNA R1b-M269 is also found in many steppe individuals who have EHG and/or CHG ancestry, but lack WHG ancestry.

Ryan said...

"Y-DNA R1b-M269 is also found in many steppe individuals who have EHG and/or CHG ancestry, but lack WHG ancestry."

EHG is 50% WHG. None of those individuals lack WHG ancestry. That's my point.

"The empirical link...."

Yes. And the link between R1b-M269's ancestors to WHG and not the steppe is just as strong. Hence why I'm arguing that R1b-M269 came from WHG - because its ancestors did. And to be clear - it's really R1b-P310 that's common in Western Europe, and that excludes Yamnaya's R1b-Z2103, which is most common among non-IE Uralic and Altaic speaking populations (though older IE populations may have been assimilated).

So there's two scenarios supported by the data here IMHO.

Scenario 1: R1b-P310 arrived in non-Iberian Bell Beakers from a steppe IE population related to Yamnaya. These IE speakers came to complete dominate the Bell Beaker DNA on the Y-chromosome but ended up speaking the Bell Beaker's language and adopting their culture.

What's unexplained:

Both the Bell Beaker culture and early IE seem to have been transmitted mostly by men. So how would Yamnaya completely wipe out the Bell Beakers' male side of the gene pool but adopt the culture of Bell Beaker women, particularly when the expansion of both Bell Beaker culture and IE cultures seems to have been male mediated? This is the crux of why I think scenario 2 is more likely - I doubt that their would be cultural and linguistic continuity of the Y-chromosome pool had such a sudden and complete turnover.

The IE-speakers that Bell Beakers actually interacted with were Corded Ware Culture too, and they seem overwhelmingly dominated by R1a. Will future samples find groups of CWC that are predominantly R1b? Will they find a CWC male that is R1b-P310 that pre-dates any interaction with the Bell Beakers? That would be the smoking gun.

Scenario 2: The resurgence of WHG ancestry and R1b and I2 frequencies that occurred throughout Europe during the Middle Neolithic also affected people a little bit further east on the steppe. So the R1b-Z2103 in Yamnaya is actually not via EHG, but more directly from WHG, probably somewhere in the Danube basin. These men assimilate into Yamnaya and coexist alongside males with R1a, Q, and older R1b clades from EHG.

What's unexplained:

What is the full story of the WHG resurgence? Will we find P310 from a male Bell Beaker who lacks steppe admixture? That would be the smoking gun.

"And, WHG ancestry is widespread in populations such as first wave Neolithic people and Mesolithic populations, in which the derived Y-DNA R1b (e.g. R1b-M269) is completely absent."

Finding only highly derived R1b is a sign that you're looking at a population that comes after the bottleneck/expansion of R1b, not before. An ancestral population should have the greater diversity.

Consider this - we only have 7 Y-chromosomes for Iberian Bell Beakers. Of those 7, 2 were R1b, 3 were I2a2 or I2a2a, 1 was I2a1a1, and 1 was G2.

Of those two R1b Iberian Bell Beakers, all we know is that one was R1b1a(xL23), and the other was R1b1. They weren't able to test either of those for R1b-M269, and only the first one were they able to exclude any mutations downstream from R1b1a. We don't actually know that R1b-M269 is absent, even for that small of a sample.

If Iberian Bell Beakers are the source of R1b-P310, then one would expect R1b to be pretty diverse in their homeland. If that's the case we may need a larger sample size too (or to catch a P310 male outside of Iberia but prior to mixing with steppe peoples).

What does seem to be known already though is that long before Bell Beakers expanded outside of Iberia, WHG had almost obliterated any first farmer Y-DNA they had. And that seems to have been happening independently in Neolithic populations in Britain and Germany too.