search this blog

Sunday, February 5, 2017

Women on the move


From the jungle known as the comments section:

First there was the creation of a new way of life north of the Caucasus, a mobile form of pastoralism herding animals which had been domesticated in the near east and the horse which was domesticated somewhere on the Eurasian steppe. Once this new way of life had been developed, it had a tremendous expansionary potential due to the vast amount of land which was suitable for it. This is why polygamy was a good strategy for these pastoralists because, as they competed with one another to build the biggest herds and control the biggest territories, it allowed for a rapid expansion of their family groups. This is the context within which there was a need to bring in additional women from outside. The pastoralists in turn would have been able to offer the families of their Caucasus farmer wives a good bride price for them.

See also...

Women on the Move. The DNA Evidence for Female Mobility and Exogamy in Prehistory

On the Caucasus as the PIE Urheimat

90 comments:

Grey said...

It would be interesting to see a study of male/female dna from the Fulani, Hausa, Nuer, Tuareg etc as if this pattern was normal for mobile raidersthis might make it easier to accept that mobile raiders will automatically become autosomally like the people they raid

http://ukcommentators.blogspot.co.uk/2006/11/sudan-incident-1909.html

capra internetensis said...

Obviously the answer is in Herodotus. The tale of the origin of the Sarmatians is a memory of a time much further back. Amazons from south of the Black Sea conquered the steppe and intermarried with the native Para-Uralic-speaking pastoralists ('Scythians'), but their descendants ended up speaking the 'Scythian' language with a Caucasian twist - PIE.

;)

I was just thinking that this new study hasn't really added anything as far the modern European R1 origins go, but actually it does raise a question. Where did R1a-M417 come from? We now have Khvalynsk I0433 far to the east, Karelia_HG in the northwest, and Ukraine_N1 in the southwest, but they are all R1a1*-M459! But there is R1a-M17 in Lokomotiv predating Khvalynsk. Was it still east of the Urals?

Davidski said...

@Capra

But there is R1a-M17 in Lokomotiv predating Khvalynsk. Was it still east of the Urals?

I know you pride yourself on being objective, but clearly, like a lot of people who post here, you need to work on being realistic.

Why would European R1a-M417 come from east of the Urals at such a late stage? Why wouldn't it come from the western steppe, along with the Corded Ware expansion, from an archaeological group like Dnieper Donets, Sredny Stog or Pontic Yamnaya, that has already been suspected of being ancestral to Corded Ware?

Keep in mind that Ukraine N is from a Dnieper Donets burial, and fits the bill as ancestral to the most steppe-shifted Corded Ware individuals.

So that comment of yours, was that a brain fart or were you just trolling, or what?

capra internetensis said...

R1a-M17 spread from the western steppe to Lake Baikal in the Mesolithic (why? how?) yet we find nothing but M459* in Europe?

I don't know why it would come from the east at this time. It doesn't make a heck of a lot of sense either way.

Nirjhar007 said...

I don't think Daves theory is totally discarded yet , but as Rob points out , it is not very attractive given current assessment of the situation .

The final pov on the PIE question in all of this, is again from where R1a M417 and Rb-M269+ emerges from and they were present to which degree throughout Eurasia chronologically . I know that in few months a massive advancement to that question will happen .

But at the end it is something, that cant be changed is that, we have no hard data to suggest which language was spoken in S Russia in those pre-historical periods , not even around 2nd millennium bc by any means , Scythian related related groups are the first attestations there .

Anything is still possible genetics wise , please stay calm as possible ;) .

Davidski said...

R1a-M17 spread from the western steppe to Lake Baikal in the Mesolithic (why? how?)

Why the western steppe? Why not somewhere like the Urals?

yet we find nothing but M459* in Europe?

Yeah, totally weird, despite all those samples we have to work with from Dnieper Donets and Sredny Stog.

capra internetensis said...

True, could just be sample bias and the main area was somewhere off to the east. But then it picks up CHG and appears in the west. Ugh, needs more ancient DNA.

Samuel Andrews said...

Other factors besides female mobility can explain higher mtDNA diversity than Y DNA diversity. So there's no real evidence in ancient mtDNA of female mobility. It's pretty much impossible to find evidence or lack of evidence.

Davidski said...

It's probably not all that impossible to find evidence of migration from one place to another based on high resolution mtDNA sequences, especially when coupled with isotopic data.

Come on, as I just told Capra, being objective is good, but you also gotta be realistic.

Amanda S said...

I've been doing a bit of reality checking on what I wrote against some of the archaeological literature. I found a paper called "The Social Structure of the Neolithic Population in the Pontic Steppe" by Nadezhda Kotova. This is an analysis of some graveyard sites in the Don and Dnieper valley dating from around 5,000 BC.
The picture it paints is of a society that has only minimally adopted aspects of farming technology; mainly cattle breeding but is still mainly dependent on wild resources for its sustenance. Although the “hunter gatherer” and “farmer” or “pastoralist” descriptions are useful categories for thinking about human societies, they can also be deceptive and over-simplifying. At this stage the society described is not really a hunter gatherer one or a pastoralist one. The writer of the paper describes a social system based around male lineage clans discernable in their burial practices.
One would think this was likely a local hunter gatherer descended society in the process of transition rather than a highly expansionary one at this stage. It raises the question as to what were the cultural elements and innovations that enabled the nascent Steppe societies to expand and when did this happen.

Davidski said...

It raises the question as to what were the cultural elements and innovations that enabled the nascent Steppe societies to expand and when did this happen.

Cultural appropriation from the farmers that they were getting their brides from?

Amanda S said...

Cultural appropriation from the farmers that they were getting their brides from?

Well, maybe. According to David Anthony, the initial spread of the farmer technologies of cattle, pottery and grain into the Pontic Steppe came from the Cucuteni-Tripolye culture of South Eastern Europe, not from the Caucasus.

I need to read more!

Maju said...

This is in answer to David, from the previous thread (Latvian and Ukrainan ancient genomes). I'll post in several separate comments to prevent the issue of max. characters, which I believe it is related, along with the captcha, with the bad gateway errors I got.

1st.-

"so the most sensible thing we can do is to focus on its PCA results"

PCA is not the best tool in any case: it's just an approximative very visual tool but we should never forget that the sampling strategy is crucial to the results. In this case we're getting one dimension clogged by the West Asia - Europe axis, so all we have left is a single dimension: the HG-EEF axis approximately (rather Kurgan-LN maybe?), that's not good enough to make any assessment. Maybe we'd get a better "feel" with Europe-only samples but in any case it's just a PCA projected on modern populations, with all the limitations that such technique implies; Alberto's approach is very different and we should not ignore it, because it gives us a different point of view: one based on ancient "ancestral" populations.

"It suggests that Corded Ware is derived from Ukraine N"...

It's possible that there is some truth to this but I wouldn't run ahead of the data. In any case it's crucial to properly understand the COMPLEX paleo-history of the region culminating in Corded Ware from an archaeological viewpoint: what happened in both Ukraine and Poland-East Germany in the early Kurgan period? Ukraine goes from DD to a very complex mosaic: (1) the critical Sredny-Stog II layer is a mix of both and even some Cucuteni and Baden influences in the case of Kiev area, (2) Yamna takes over, except for the NW, which falls to Luboń and Globular Amphorae (Poland-based cultures), with some overlap near Kiev, (3) Catcacombs culture appears, another critical element of unclear (but Kurganic) roots: may be a local re-shaping of Yamna (but Yamna persists farther East) or it may even be influenced by Maykop, in any case this culture is also found in Cuyavia at the very genesis of Corded Ware, overlapping Globular Amphorae, and it's suspected that the hybridation of both is what triggers Corded Ware (with core in Cuyavia). I'm not going to detail the sequence Baalberge-various groups-Luboń-Globular Amphorae in Poland and East Germany, with expansion to West Ukraine, but that's also Kurganic at the root (Baalberge) although maybe of mixed nature (Funnelbeaker influences from Denmark or elsewhere in the Northern coasts) and later also with Baden lesser influence from the south (the "taming" of early West Indoeuropeans?). So, from the viewpoint of archaeological based prehistory there's a wide range of areas that could be crucial in the formation of Corded Ware and the seminal episode of R1a1-European associated to it: from the Elbe to the Don and beyond but most likely somewhere between the Vistula and the Dniepr, most of which was non-Yamna, even if culturally related at some points. So there's a good likelihood that the "founding fathers" of CW were Indoeuropeanized peoples from some other background, something between EEFs and "barbarians from the Baltic".

Davidski said...

@Amanda

Well, maybe. According to David Anthony, the initial spread of the farmer technologies of cattle, pottery and grain into the Pontic Steppe came from the Cucuteni-Tripolye culture of South Eastern Europe, not from the Caucasus.

Yamnaya does show that type of western ancestry. It doesn't usually show up in ADMIXTURE, but it does in models based on formal stats, and even in my nMonte tests.

Some of the brides, and a lot of the innovations, could have come from CT.

Alberto said...

But regarding this "West Asian" mtDNA, we just got a new paper that looked at 12 samples from Ukraine Bronze Age, spanning from pre-Yamnaya, Yamnaya and post-Yamnaya periods. And they're all U, U5, U5a and a couple of C4.

And not long ago we got mtDNA from the Botai culture (who were basically hunter-gatherers, from Kazakhstan) and their results were all "West Asian" IIRC.

Could be just luck, but that's what we have right now.

Davidski said...

@Maju

PCA is a very good tool if used properly and interpreted correctly. This is what I've done, and I've tested the PCA results with formal analyses, as you'll soon see.

Don't forget also that the case I'm making is fully in line with the conclusions of the people who published the data and indeed the current consensus.

So the onus is not really on me to prove anything, since you're the one arguing against the tide here. But what the hey.

So there's a good likelihood that the "founding fathers" of CW were Indoeuropeanized peoples from some other background, something between EEFs and "barbarians from the Baltic".

This doesn't work. I'll show you why later today. You need to make an effort to understand it though.

Maju said...

2nd.-

"How did Caucasus Hunter-Gatherer ancestry get there? Amanda S explains here..."

I thought that was one of the absurd ideas that made the discussion "muddy" (to put it politely). IMO it's a mere attempt at whitewashing what was surely slavery and rape by rather brutal conquerors. In any case, the Fulani are a bad model because for all their recent expansion within the Fula Jihad of the 18th century, they have made near zero impact in the neighboring/subjugated populations in terms genetic, if the Fulani would be the model, the result would just be persistence of pre-existing farmer populations and marginality of the herder-conquerors, and that, we all agree, is not what we see in Late Chalcolithic Europe and aftershocks. Another, more European and steppary example could be the Magyars, who left a genetic legacy that is nowhere to be seen, but their linguistic and ethnogenetic impact is clearly very strong however. I think that cases, surely not as extreme as that of the Magyars (or say South Slavs, same issue) but to some extent comparable, happened once and again in this process of Indoeuropeanization of Europe, with mixed populations, maybe just assimilated sometimes, being ethno-linguistically transformed into Indoeuropeans by mere elite dominance, conquest or, as the Slavs did, systematic incorporation to their ranks of almost every single defeated population/individual (this "adoption" process is very typical of many tribal societies anyhow and in the even of expansion may lead to total dilution of the original genetic pool of the source population).

There are no easy answers, that much we have learned in the last years I hope.

"Corded Ware is very similar to Yamnaya, and that it might even be an Yamnaya offshoot."

But is it EHG or is it SHG? These are different things (even if both carry "extra ANE") and while Yamna and, I believe, CW are EHG+IranNeol (or +CHG if you wish, IMO not), UkraineNeol seems rather strongly leaning to SHG. So there's an issue here that PCA won't solve.

Also Yamna and its precursor Khvalynsk were dominated by R1b-Volga, while CW is by R1a1 and there is where the issue of secondary populations, maybe in NW Ukraine or Poland comes handy, because R1a1-Europe seems to have primarily expanded from either of those areas, areas that are characteristically non-Yamna. Of course there's a chance that the lineage was picked from Catcacombs in a founder effect we just cannot discern with the data available: we'd need some hint pointing in that direction to be able to make the claim with any credibility and so far we have nothing.

Maju said...

"... you shouldn't expect any surprises from Central Poland. Expect Y-HG I, mostly I2, from prior to Corded Ware, and R1a only after Corded Ware. This isn't speculation on my part."

Unless you provide a source or give me a more specifically clear declaration that you know that for a fact (i.e. from privy access to yet-unpublished data), I don't see any reason not to consider that "speculation on your part", sorry.

Samuel Andrews said...

Just found two more Steppe mtDNA lineages; R1 and H11a.

Samuel Andrews said...

@Maju,
"In any case it's crucial to properly understand the COMPLEX paleo-history of the region culminating in Corded Ware from an archaeological viewpoint:...."

Like last year and the year before you look to archaeology and modern DNA to form opinions on mysteries which have already been solved by ancient DNA. You were wrong two years ago and you're wrong now.

Maju said...

"This doesn't work. I'll show you why later today. You need to make an effort to understand it though."

I see how it "doesn't work": because CW are or seem to be too much strongly Yamna-like in the autosomal DNA (although they do show some EEF-like admixture it's probably not enough). But what I say is that we just don't know how exactly things paved out in Ukraine, Poland, etc. We do not have neither Y-DNA nor in most cases nDNA from those areas to assess the issue yet. What if Baalberge-Luboń-GA in Cuyavia was more strongly Yamna-like than their relatives at the Elbe? What if the same scenario took place in the complex landscape of Sredny-Stog II Ukraine-Don? In those cases there is a lot of time for a lot of things to happen, really, including that a Indoeuropean clan was co-opted by originally non-IE men, who adopted (then or earlier) IE language and most of the culture. So I'd try to keep my mind open, really.

It can still be that R1a-CW comes from Catacombs (what may be attributed to Yamna or Maykop or DD/Cucuteni underdog persistence or whatever), I'm not excluding that line but I just don't see any evidence supporting it yet and by evidence I mean some sort of ancient DNA from Catacombs that tells us that it's a good trail (best would be direct strong R1a evidence, of course).

"PCA is a very good tool if used properly and interpreted correctly."

It's extremely approximative, of almost no use in our case, where we're trying to discern EHG vs SHG influences, much less with all those useless West Asian samples clogging one dimension. Sorry but PCA analytic power is very limited and that's hardly a matter of debate. Most of the time using a bidimensional PCA is the equivalent (at best) to a K=3-4 depth in ADMIXTURE analysis, typically not good enough. An effectively one-dimensional PCA such as the one discussed here is equivalent to a K=2, what is as bad as it gets.

Maju said...

@Samuel: I'd gladly reply to you if I could find out what the heck you're talking about. I just can't.

Davidski said...

It's extremely approximative, of almost no use in our case, where we're trying to discern EHG vs SHG influences.

But it's very easy to tell apart EHG from SHG on my PCA, as well as Ukraine HG/N, and thus their influences.

Take another look at the first PCA, and answer honestly, does Latvia LN look like a mixture of Ukraine HG/N and CHG as per formal models, or not? And do you think this is a coincidence?

Samuel Andrews said...

@Maju,

You mostly analyse archaeology and modern DNA in order to create your opinion of the genetic origins of modern Europeans. Everyone else mostly analyses ancient DNA. Ancient DNA is better than modern DNA and archaeology.

Maju said...

OK, David, fair enough: ancient Latvians in your PCA look like "extreme Motala", "extreme SHG" with no obvious tendency towards CHG or IranN. I don't know how that happens but I don't know how the PCA is a better explanation here either. It's possible that the limited amount of "ancestral" populations used by Alberto is causing some issues but it's also possible that the "clogging" of the PCA by so many diverse populations is not allowing LatN to place themselves in the "correct" coordinates (whichever those are). Maybe you should run a simple PCA with just CHG/IranNeol, Motala, EHG and LatNeol and see what happens.

In any case I will concede the doubt in this particular case. But we're getting a bit lost: your original comment was: Ukraine N is a low quality sample, so the most sensible thing we can do is to focus on its PCA results in the most significant dimensions.

So how did we end up jumping from UkraineN to LatviaN? UkrN is, per Alberto's work, 66% SHG, 24% EHG and just 9% CHG. We can go without the CHG issue although I would expect Neolithic Ukraine to have some Caucasus or Zagros farmer influence maybe. What I understood you meant to redefine was UkrN as not being so strongly SHG or even at all. Yet what does your PCA say here? Well, it's pretty close: between EHG and SHG, maybe it seems to tend a bit more to EHG but that can easily be attributed to the quirks of PCA and may well be a false call.

Davidski said...

There's nothing quirky about that PCA. Latvia_HG, Nordic_HG (SHG), Ukraine HG/N and Eastern HG are all exactly where they should be in line with their levels of Ancient North Eurasian admixture.

Latvia LN, Corded Ware and Yamnaya are also exactly where they should be based on what we know about their ancestry from a lot of different tests.

Azarov Dmitry said...

@ capra internetensis
I was just thinking that this new study hasn't really added anything as far the modern European R1 origins go, but actually it does raise a question. Where did R1a-M417 come from? We now have Khvalynsk I0433 far to the east, Karelia_HG in the northwest, and Ukraine_N1 in the southwest, but they are all R1a1*-M459!


I have already explained everything in my earlier comments from August 16, 2016 so I just repost them
================================================================
Azarov Dmitry said...
@Nirjhar007
Y-HGs J, G, L, H, and R1b are all possible for Maikop.

You forget R1a-M417. If its there, it will solve the CWC R1a situation , although I am sure you think its already solved..


I guess Maikop guys will be mostly carriers of R1a-M417 subclades (something like R1a-Z645) while majority of CWC R1a guys will be from R1a-YP1272 subclades. For a while the best model for R1a folks expansion is a 3-wave migration from the Iranian Plateau:

1-st wave – R1a-YP1272 folks, migration rout: Iranian Plateau->Caucasus Mountains, affiliated cultures: Sredny Stog, Cernavodă, Ezero, Yamnaya (Western part), Baden, CWC, Fatyanovo, Abashevo;

2-nd wave – R1a-M417 folks, migration rout: Iranian Plateau -> Mesopotamia -> Caucasus Mountains, affiliated cultures: Hassuna, Halaf (transition), Maikop, Catacomb, CWC, Trialeti, Andronovo, Srubna, Trzciniec, Tumulus. I believe R1a-M417 folks (Maikop -> Catacomb cultures) pushed R1a-YP1272 and R1b-Z2103 folks from territory of Yamnaya culture.

3-rd wave – R1a-YP1051 folks, migration rout: Iranian Plateau -> Mesopotamia ->Anatolia -> Balkans, affiliated cultures: Ezero and Western Anatolian cultures.

Map of migration routs for R1a folks from the Iranian Plateau:

http://s019.radikal.ru/i606/1608/5b/a00d11bfa19a.jpg

http://s06.radikal.ru/i179/1608/50/b6ff13b5574f.jpg
August 16, 2016 at 12:58 PM

Maju said...

@Samuel: "Ancient DNA is better than modern DNA and archaeology".

It is not "better", it's just another data set, sadly too often not good enough (because of limited sampling or geographic coverage, sometimes also because quality issues). I do use ancient DNA anyhow but I do not "worship" it, as you seem to do, as the only possible type of evidence, furthermore I have actually used often ancient mtDNA particularly, which is typically telling the same story as autosomal DNA (not always but in most cases it does, while Y-DNA holds a much weaker connection at least some times, being more strongly subject to distorting founder effects) and has the unique advantage of a much better geographical and chronological coverage than autosomal or Y-DNA, way too often available only for very specific sites or geographies/chronologies.

So I try to look at the whole picture: modern genetics and what inferring from it allows us to teconstruct a plausible paleo-history (sometimes very interesting, like the case of the Y-DNA K2 plus mtDNA N/R expansion from SE Asia, probably after the Toba event) of even the "out of Africa" episode, which archaeology has confirmed pretty well but for dates much older than those proposed by "molecular clock" enthusiasts. I look at archaeology (fundamental, how would you know what is even Corded Ware or Yamna without it?), I look at ancient DNA from all available viewpoints (nDNA, mtDNA, Y-DNA), etc. And it is looking at all the pieces of the puzzle what makes it work, while looking at only some very specific data (what is the very definition of cherry-picking the evidence BTW) will almost unavoidable produce errors in the analysis.

What happens when the data is contradictory? For example the issue of where in Africa is the greatest (basal or general) diversity for each of the DNA "fractions": mtDNA (Sudan-Ethiopia), Y-DNA (Central-West Africa), nDNA (Central-Southern Africa). Well, look at the data again first of all and try to conciliate them to a common element, or maybe argue that you think that this type of data (mtDNA in my book, largely because archaeology supports it) looks much better because the others can be explained by either archaic admixture (Y-DNA A0 and A00) or that plus other quirks (nDNA). But in the end it will be an educated guess, which you can respectfully disagree with, until someone invents the time machine and we can directly film the key events as they actually happened.

Maju said...

@Alberto: re. what you say above, it may be of interest something that I think Nirjhar mentioned once which is the ill-known Yangelskaya or Angelskaya culture (?) of the Ural river, technically in Asia but very close to where the Samara-Khvalynsk-Yamna core of Kurganism would later show up. These seem to be (on archaeology) a mix of "aborigins" (EEF?) and Zagros Neolithic, which was argued to have arrived there via the Eastern Caucasus corridor. I suspect that is an important culture in order to understand why there is so much "West Asian" towards the East and so little elsewhere in Eastern Europe initially at the Neolithic. Another route that the Neolithic West Asians took was to the Central Asian "oasis" region (Uzbekistan and surroundings) but I find hard to imagine they reached the area of Botai that way (uncertain though).

Maju said...

@David: your PCA does not have Ma1 or AG (i.e. "ANE", which I find hard to believe you are still even mentioning as informative but whatever) and even if it had it, it would hardly constitute a polarity in that kind of PCA (EHG or an approximation based on modern populations would instead), so whatever: I just don't see it nor I see how it matters re. the SHG vs EHG issue (sure: "ANE" is one of the traits they differ on but it's surely not the only one).

Amanda S said...

I thought that was one of the absurd ideas that made the discussion "muddy" (to put it politely). IMO it's a mere attempt at whitewashing what was surely slavery and rape by rather brutal conquerors

Whilst it's hard to imagine that the expansion or to use the more human term, invasion of the Steppe people into areas settled by European Neolithic farmers was not characterised by a great deal of violence and conflict (when has such a big demographic change not been), I don't see why the relationship between the North Caucasus farmers and the Steppe people could not have been a mutually beneficial one. The Steppe people did not covet their land and there must have been a healthy trade between them, metals from the Caucasus in exchange for the products of the Steppe. I'm only putting forward a scenario which could explain the observed findings of the mixed origins of this population and its sexual asymmetry.

Rob said...

@ AmandaS

The invasion, migration or whatever term of steppe peoples into north-central Europe seems to have followed a demographic decline in mid-late Neolithic Europe, although this might have been regionally varied. It also occurred along specific niches, sometimes even marginal land (perhaps a sign of their flexibility).

Back to the Caucasus - steppe interaction, with regard to Kotova's excellent analyses and Shishlina's work (which you appear to be indirectly referencing), it should first be noted that the populations in the north Caucasus were themselves pastoralists, or at least partially so, which explains why they might have moved north onto the open steppe to begin with.
In regard to the matter of the steppe populations themselves, there appears to have been a change from burials where both men and women were adorned with gifts, to mostly. virtually exlcusively males in the Eneolithic and subsequent Yamnaya period. And although the social structure was characterised by exogamic clans, it is curious as to why between the Ukraine_HG and Ukraine_Neol, and Samara _HG and Samara_Eneolithic, the genetic shift toward CHG is is very minimal over several thousands of years, which makes the changes between Khvalynsk and Yamnaya seem rather sudden.
This means that the social shift which created all-powerful shifts was rather sudden, or the alternate possibility is that there was a migration/ invasion from the region of Majkop. I favour the second, because this happened not only in the broader Yamnaya region, but also in far away Afansievo, where no such preconditions (thousands of years of farmer- incipient pastoralists interactions occurred, according to present evidence).

Matt said...

Re: LN1 and PCA, running in nMonte with a merge of the Globe10 and Days of High Adventure PCA, (9 dimensions from DoHA, 10 from Globe10) I found:
Latvia_LN1:ZVEJ28 - Kotias 40.55, Latvia_HG 34.15, Samara_HG 25.3, Barcin_Neolithic 0, Iberia_Mesolithic 0, Loschbour 0, Ukraine_N1 0 - distance% = 2.7119 %

No Ukraine N1 preferred there.

With G10 alone:

Latvia_LN1 - Kotias 42.1, Latvia_HG 31.25, Samara_HG 26.65, Barcin_Neolithic 0, Iberia_Mesolithic 0, Loschbour 0 Ukraine_N1 0 - distance% = 2.2073 %

Could be due to artefacts in placement on the PCA of the low coverage Ukrainian samples maybe?

Though based on the sequence of these samples, Latvia_HG was superceded by Latvian MN, which was more WHG, and Latvian MN2, essentially EHG, so I don't know that you would have a good historical reason to expect continuity from Latvia_HG themselves. If MN2 had replaced Latvian_HG they could have survived even more strongly in LN1 than you might expect.

Yamnaya_Samara under DoHA+G10 merge: Samara_HG 50.65, Kotias 38.2, Iberia_Mesolithic 6, Latvia_HG 5.15, Barcin_Neolithic 0, Loschbour 0, Ukraine_N1 0 - distance% = 1.0925 %

(I think DoHA or using it with Globe10 should be generally superior for combinations of West Eurasian ancients, as it actually recreates structure between modern Europeans while G10 collapses almost everything into a single cline. But it does seem to have a flaw in picking up Iberia_Mesolithic and Iberian Neolithic populations in steppe and steppe-like when nMonte allows them as a population; some blur in discriminating Iberia vs Steppe HG in higher dimensions. I think this may be a bug in the PCA that you might want to look at @Davidski if you ever want to rebuild this PCA.).

with just G10: Samara_HG 49.75, Kotias 40.05, Latvia_HG 10.2, Barcin_Neolithic 0, Iberia_Mesolithic 0, Loschbour 0, Ukraine_N1 0 - distance% = 0.4429 %

Amount of Kotias ancestry in each seems comparable (not diluted in LN1).

*For example of how DoHA is better at recreating the real structure of European populations, look at this sequence of four MDS, left to right : http://i.imgur.com/zHlqlbI.png

The first is based on fst, so recreates structure, but is bloated by some high particular to one or two populations (e.g. Orcadian, Ashkenazi) and places diverse populations close to others (e.g. Spanish). The second is based on a PCA Davidski ran of only Europeans, so should be a relatively pure measure of intra-European differentiation, uninfluenced by compromising to fit non-European populations on.

Third is based on DoHA and is a good approximation of the first two, while the fourth based on Globe10 is the odd one out and collapses everything in North-Central Europe into a single cline.

Davidski said...

If MN2 had replaced Latvian_HG they could have survived even more strongly in LN1 than you might expect.

Yeah, but then you'd need a migration of a very CHG-heavy steppe population to the Baltic, which then just happened to create a Yamnaya-like effect, just slightly more westerly.

Not very parsimonious and not in the ancient DNA record.

Rob said...

Well that's what appears to have happened in the Balkans, so why not Poland, the steppe and Baltic : CHG admixing into different local substrates, yet all containing some EHG; giving the appearance of a "Yamnaya migration". Of course some traditional pre-Yamnaya steppe EHG hitched along for the ride

Only Majkop aDNA will confirm or deny this

Rob said...

That paper Romulius posted a couple of days ago- http://s3.amazonaws.com/academia.edu.documents/11520495/Interaction%20between%20hunter%20-%20gatherers%20and%20farmers%20during%20the%20Neolithic%20in%20Poland.docx?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=1486274376&Signature=Azgp2XNroHeZ9Il%2BPrWHi9KwuOM%3D&response-content-disposition=attachment%3B%20filename%3DInteraction_between_hunter_-_gatherers_a.docx

Fig 3: Poland was a rich hunter-gather land well into the Bronze Age. "Latvia HG" type groups existed there too, imo.

Matt said...

Yeah, it doesn't really look like that at the moment - does seem like you would need some mostly pure CHG herders around in the steppe-forest steppe who did what Amanda talked about ("a rapid expansion of their family groups") only maybe with males and not a female bias. No evidence any such groups existed (genetic, I know it doesn't; archaeological, I don't know enough about the archaeology).

Nirjhar007 said...

I do not agree to the suggestion that aDNA is better than modern DNA , both actually works as a complementary to each other .

Rob said...

Fair enough
A pure CHG group doesn't seem realistic

Maju said...

@Amanda: Let's assume your working hypothesis, why would the result be so one-sided then? Wouldn't it'd be logical that the interaction would produce a bilateral exchange and that the North Caucasus would not just become part of the Khvalynsk culture as it happened in reality but would retain its own specificity?

In my opinion there was indeed a North Caucasus alliance of some sort between early Indoeuropeans (probably of the Anatolian linguistic branch) and aborigines (probably of the NE Caucasian and Hurro-Urartean branch) but this I deduce from what happened south of the Caucasus in the Anatolia-Armenia-Kurdistan region, and I'm unsure on how it works for the North of the region, other than Maykop and Kura-Araxes being clearly related. It's perfectly possible that slavery and collaboration happened simultaneously with the same ethnic group, much as we see in historical slavery episodes in the Slavic countries (Slavs were both victims and perpetrators of the slave trade) or Africa (roughly the same thing). In any case I strongly feel that for the male bias to happen there must be a sustained one-sided conquest of one population (the "founding fathers") on another (partly the "founding mothers"). There are circumstances in which, with lots of time and a clear demographic imbalance, this is not needed (the Uralic Europeanization by the mothers' side, for example), but in these cases we don't seem to have enough time nor is either obvious the sort of density imbalance we can find further north.

"The Steppe people did not covet their land"...

The North Caucasus is part of the steppe, just a somewhat milder weather but steppe in any case. So for me there is no question on why would pastoralists seek to establish themselves there: good pastures (among other secondary reasons). It's not at all like the relation with Uralics of the Taiga, who occupied a distinct eco-niche.

"metals from the Caucasus in exchange for the products of the Steppe."

In exchange of gold from the Urals maybe (easy to identify because it has a strong fraction of platinum, for what I've read), because products specific of the steppe (leather, horses, cattle?) would exist most of the time in both regions, because the North Caucasus is part of the steppe region.

"I'm only putting forward a scenario which could explain the observed findings of the mixed origins of this population and its sexual asymmetry".

Do we even have Y-DNA from Maykop? We don't know if there was any "sexual asymmetry" (the expression sounds a bit weird but let's ignore that). It seems as you guys are talking about something that (allegedly) "objectively happened" but when you look at it, there is nothing: no data, no nothing. So I wonder why all this speculation on top of another speculation whose factual reasons are nowhere, really. Is it all about some "CHG" that appears in autosomal analyses? How from just that can you jump to such an elaborate speculation about "gender imbalance"?

Notice I'm not blaming you particularly, Amanda, there are more accomplices in this, really, seems to me like groupthink: when people in charge of something reinforce the group's opinions, often in the wrong direction leading to catastrophic results. Beware of that (that's why generally democracy works better than dictatorship: you get feedback from the exterior of the power rings thanks to freedom of speech and that may have saved the day more than once).

Davidski said...

It seems as you guys are talking about something that (allegedly) "objectively happened" but when you look at it, there is nothing: no data, no nothing.

Don't exaggerate. There's plenty of relevant data available.

Shaikorth said...



@Maju

We know from X-chromosome inheritance that there is male-biased steppe ancestry in CEU (NW Europe):
http://biorxiv.org/content/early/2016/09/30/078360
...and furthermore we know that most Europeans (everyone in 1000genomes at the very least) show a similar degree of male-biased demographic history as CEU. This means the steppe ancestry infusion was a major event, a male-biased one at that, and perhaps the only one of its kind in Europe.



Example: any eastern male-biased ancestry relative to what CEU have will make a population's X-chromosome significantly more EEF and WHG than its autosome. Sardinian late Indo-Europeanization was male-biased since they are the only population showing it relative to CEU.

http://biorxiv.org/content/early/2016/12/07/092148
https://postimg.org/image/c62wb3tgp/

Matt said...

Btw, don't know if anyone was interested, but combining the DoHA and Globe10 PCAs and running Nmonte values for recent Europeans produced this set up:

Calc: BarcinN, Kotias, Iberia_Mesolithic, Loschbour, Latvia_HG, Ukraine_N1, Samara_HG

Table: http://i.imgur.com/61xETRf.png

Put through MDS: http://i.imgur.com/DWzVg4B.png / Clustering: http://i.imgur.com/vUQTbKd.png

Even merging in the Globe10, there is still a preference for the local HGs in data using Days of High Adventure PCA and the two Neolithic approximates. So even if it's not likely that local HGs actually contributed, it looks like the combination of different HG sources may approximate them having contributed, in an odd sort of way (additional new HG from East and West evens out to a similar thing?), in these dimensions (whether they tell the full story or not).

Nirjhar007 said...

Anybody experimented yet, how S Asians will perform, in modelling, having the new samples?.

capra internetensis said...

@Matt

Isn't that X chromosome data for Steppe + EEF though? Is there anything comparable for CHG + EHG in Yamnaya?

@Azarov Dmitry

That does not explain the M198 at Lake Baikal, which was my point.

I did see your original post, but frankly I did not find it more convincing than anyone else's speculation. Not that I rule out Maykop, just there is not much to support it right now.

Matt said...

@ Capra, I believe you want Shaikorth, but IRC, no.

capra internetensis said...

@Matt

doh!

Is Iberia_Mesolithic the two La Branas?

It does make sense geographically albeit kind of noisy; I guess in cases like northern Italy it's saying we are missing the appropriate reference?

Matt said...

@ Capra, I'd assume so; they're just labeled as Iberia_Mesolithic in the datasets.

Btw, other Mesolithic era HG samples, I could've run with the Hungarian_KO1 and Motala. At least for the merge, as the others (BerryAuBac, Chaudardes) didn't really cross up between Globe10 and Days of High Adventure.

I was really impressed with the geographical pattern, in these sheets (the merge, but also the pure Days of High Adventure) where populations tend to come out with lots of local HG (albeit with particularly Samara noise in the merge)... but then the local ancients from the same region really don't fit (e.g. Iberia_Chal takes on Loschbour, not Iberia_M, Corded_Ware Germany takes Latvia_HG but also Iberia_M, Sintashta for some reason takes on Iberia_M etc.) - http://i.imgur.com/wO8WTt2.png.

The ancients do come out with more or less the exact expected amount of Barcin_N and Kotias (e.g. 74% Barcin in Iberia_Chal, 38% Kotias in Yamnaya, more or less intermediate the two in Bell Beaker Germany). Puzzling.

Grey said...

Maju

"In any case, the Fulani are a bad model "

Maybe so but in any case if you read a lot of military history mounted raiding is historically attested hundreds of times - happening today in places like Sudan - so it wouldn't be surprising if it happened in prehistory also.

Maybe it doesn't have to be that. Friendly bride exchange might explain it if the population density was very skewed i.e. 10% bride exchange between 100 pastoralists and 500 farmers would be 10% farmer mtdna in the pastoralists but only 2% pastoralist mtdna in the farmers.

(if arithmetic wrong you get the idea)

I don't know how to calculate that quickly enough to not get bored but maybe if repeated over a long enough time frame maybe that could get the skewed result?



Personally (admittedly based mostly on military history) I think the most likely scenario is:
- farmers moved out onto the steppe as far as it was viable
- they catalyzed the HGs beyond the border into pastoralists
- friendly coexistence or only minor conflict for a long while
- over time the pastoralists got militarily stronger (horse breeding?)
- eventually they got strong enough to over-run the farmers on the flat lands
- (but not the mountains)
which if correct would mean they'd start with low levels of CHG mtdna from friendly relations or minor raiding and then get a whole bunch during the last phase.

#

It's not my favorite theory - my favorite is natural selection on metabolic rate so the hot-blooded but high maintenance jotun women got gradually selected against.

Maju said...

@Grey: the density explanation works well for far north phenomena implying a long term (emphasis: long term, not short term) unequal genetic flow via women mainly (patrilocality) northwards. This works very well for European Uralics and probably also for proto-Amerindians in NE Asia in a pre-LGM phase. It has the (rational) assumption that the more southern, the denser and that the gradient is rather abrupt, maybe 10:1 or more, plus also various layers of populations in the "matrilineage conveyor" northwards.

But when we come to the pastoralist-farmer alleged scenario we stumble on all kind of problems: (1) how clear cut was that duality? We know that Botai people were strict horse pastoralists but otherwise the populations probably used both techniques (plus hunting and fishing) depending on what worked best for them in each eco-niche and cultural premises; (2) there's absolutely no reason to imagine that strict pastoralists could have greater densities than farmers (usually also with pastoralism) anywhere, rather the opposite (in all places where farmers and herders live side by side but as separate populations, herders are the minority, farmers are the majority by a lot); (3) the North Caucasus has a milder climate than the Volga bend, just because it is farther towards the south, so densities should be (everything else equal) greater there and not the opposite.

"my favorite is natural selection on metabolic rate so the hot-blooded but high maintenance jotun women got gradually selected against."

That's a joke I hope.

Maju said...

@Matt: that analysis of yours is very interesting: it makes very difficult to argue against some sort of CHG (Iran Neolithic??) penetrating Latvia in the Neolithic. Can I use that table (with due attribution) along Alberto's?

Maju said...

@Saikorth: interesting paper indeed (thanks for mentioning) but I can't find the acronym "CEU" in all the study: it is talking of extreme male bias in Bronze Age Central Europeans (not sure which exact samples right now). The acronym "CE", which may have confused you, is used but for Central European Neolithic peoples and what they do find is massive male-biased steppe admixture in "BA" (Central European Bronze Age peoples).

I was scratching my head until I realized that, because CEU (as primarily "Anglo-Danish") are rich in R1b-Western (sub-L11) but Bronze Age Central Europeans not at all (mostly I2 and R1a).

Maju said...

@Saikorth: we can do the following exercise, while we await for more data, in order to estimate the fraction of Western IE patrilineal ancestry in each modern population: I assume that there was c.50% R1a in early Western IEs (Chalcolithic/BA Central Europe; Poland is 57% today, so it's a good guess), then what patrilineal fraction belongs to those origins in each of the following areas (mostly to the west of the core region, sorted by est. fraction)?

Norway: 51%
Germany: 32% (18-48% depending on region)
Serbia: 32%
Sweden: 32%
Denmark: 30%
Scotland: 17%
England: 9%
Italy: 8% (6-9% depending on region)
Netherlands: 8%
Switzerland: 7%
France: 6% (0-11% depending on region)
Ireland: 5%
Spain: 4% (0-19% depending on region)
Portugal: 3%
Sardinia: 2%
Wales: 2%
Corsica: 0%
Basque Country: 0%

Of course this is a rough approximation, because there can be founder effects carrying more of X1 lineage and none of X2 lineage, etc. but I think that for a general outlook it's a pretty good approximation. Notice that I only mean direct patrilineal, that autosomal ancestry via admixed intermediate populations can indeed be greater or lower or whatever.

[Note: frequencies of R1a taken from Eupedia]

Shaikorth said...

@Maju

Y-haplogroup frequencies, subject to extreme founder effects as they are, are uninformative in determining sex biased demography, you need to look at the X/autosome comparison. We don't see the bias in X/autosome relations of 1000genomes Europeans towards LBK or Loschbour for example (see the Sardinian paper). Relative to CEU that is, so if you assume CEU lacks the bias so does everyone else (except Sardinians). Since the basis of NW European genepool was established in late Bronze Age it isn't unreasonable to assume LBA sex bias hasn't vanished.

For the record, R1b was found in Bronze Age Germany and Hungary and more importantly in the Irish Rathlin samples who are the very image of BA intrusion compared to their Neolithic predecessors.

Davidski said...

Matt: that analysis of yours is very interesting: it makes very difficult to argue against some sort of CHG (Iran Neolithic??) penetrating Latvia in the Neolithic.

CHG or better yet Iran Neolithic penetrating Latvia in the Neolihic. Holy shit. Hahaha.

Maju, your are truly hopeless at this.

Maju said...

David: I'm just stating what two of your commenters, using your analysis tools, have apparently found. So less personal attacks and more humble reasoning and self-criticism, please.

Maju said...

That above was for @David, naturally.

@Saikorth: I've been looking at the Germany BA Y-DNA data and none of them, other than the BB individual (a SW-Europe-originated culture) carry R1b-Western, all the rest carry unspecific upstream lineages, which may in some cases (??) belong to R1b-Volga (haven't checked today). Anyway R1b, even including upstream paragroups, is clearly in the minority end and like 90% are R1a or I2.

I'm sure someone can come with improved techniques, especially provided enough aDNA data, to estimate the CW-like ancestry in modern Europeans, but as approximative rule of thumb, my system looks pretty good. My suspicion is that it may understimate French (whose IE-ization may have come by the hand of populations tending towards the I2 side maybe - but not using I2 because it's been everywhere around since Paleolithic, so unreliable) and overstimate Norwegians and neighbors (their R1a includes upstream lineages, too basal to be certainly IE/CW).

"Relative to CEU that is, so if you assume CEU lacks the bias"...

Why do you insist on hammering with CEU? There is no CEU in your first linked study just ancient Central Europeans.

Grey said...

Maju

"That's a joke I hope."

partly

https://en.wikipedia.org/wiki/Yaghan_people#Early_Yaghan_people

"The Yaghan ...snip. They were famed for their complete indifference to the bitter weather around Cape Horn.[8] Although they had fire and small domed shelters, they routinely went about completely naked in the frigid cold and biting wind of Tierra del Fuego. Women swam in its 48-degree-south waters hunting for shellfish.[9] They often were observed to sleep in the open, completely unsheltered and unclothed, while Europeans shivered under blankets.[5] A Chilean researcher claimed their average body temperature was warmer than a European's by at least one degree.[7]"

Unless there's an explanation for this other than high metabolic rate i think it's likely humans in the far north may have had similar metabolic adaptations to keep them warm

(which i assume required more or specific types of food)

and mtdna is connected to metabolism/heat (apparently).

(just an idea and not very likely maybe but it would still be interesting to see if the population with the most old north euro mtdna had a higher average body temperature)

Shaikorth said...

@Maju

CEU is relevant here because it represents a genepool formed by people like LBA Central Europeans. Therefore it likely has the male sex bias of these ancients, and so do other 1000genomes Europeans as per the 2nd study I linked (they all have CEU-like X-autosome ratio in their relationship to Neolithic farmers and WHG, needless to say a more male-biased eastern ancestry would skew the ratio like it did for Sardinians). Or it doesn't for some weird reason in which case other 1000genomes European samples - and probably almost all other Europeans - don't have a sex bias.

Y-haplogroup frequencies are uninformative when it comes to confirming sex-biased demography. Sardinia has high G2, I2 and R1b-V88 which are not steppe-associated (but connect to WHG and EEF) but their autosome is distinctly less EEF and WHG relative to their X when compared to CEU, or other 1000genomes Europeans who all have a CEU-like balance. In other words, while Sardinians have less eastern ancestry than other Europeans it's clearly more male-biased in their case regardless of Y-DNA.

Maju said...

@Saikorth: sorry, it seems I misunderstood you re. CEU (I basically ignored the paper on Sardinians, so I was confused, my bad). What the graph you linked to suggests is that there is a correlation between autosomal DNA and X-DNA in all sampled populations but Sardinians, so Sardinians and not CEU (nor FIN, GBR, IBS nor TSI) have some sort of gender-bias. Basically what that graphic says is that there is no gender bias in Europe (except apparently in Sardinia) if we use CEU as reference.

Did I get it right this time? I think so.

If we take that argument to its logical conclusion: mtDNA pool = nDNA pool = Y-DNA pool (except in Sardinia). Of course it's also possible that CEU has some sort of bias which should be almost exactly the same for all other European populations (except Sardinians) but that looks unlikely on first sight (why would Finns and Iberians have the same level of gender-bias unless it's approx. zero?)

"Y-haplogroup frequencies are uninformative when it comes to confirming sex-biased demography".

What I meant was to illustrate approx. patrilineal ancestry only, not overall ancestry, that belongs to autosomal DNA, naturally. And that's why I suspect that there is hidden patrilineal CW-derived influence in French: because they consistently appear (Günther & Valdiosera for instance) as more Caucasus-influenced than Iberians, about double, and that is not straightforward with my rough Y-DNA analysis (but French are particularly rich in I2, so that's probably where the answer lays). Otherwise Iberians are consistent with a no strong gender bias scenario in admixture: their Caucasus component is very small, and so is their R1a frequency, and they behave like CEU or other Europeans in the X-chr analysis you mention.

I'd rather think that it is Sardinians the ones showing gender-biased admixture of some sort, but it may well be a very ancient founder effect. A recent paper on Sardinians indeed suggested that, while their autosomal DNA is clearly EEF-like, their Y-DNA is not, and the author claimed therefore that, while the basics of the "triangular model" are correct, there must be a lot much greater nuance, in the case of Sardinians by means of picking up pre-Neolithic lineages like I2 (he also thought R1b was in that category, although I know that many here will disagree - but the lead author's surname is Chinese, so he's probably less biased than you or me on European affairs just because of that).

Shaikorth said...

"If we take that argument to its logical conclusion: mtDNA pool = nDNA pool = Y-DNA pool (except in Sardinia). Of course it's also possible that CEU has some sort of bias which should be almost exactly the same for all other European populations (except Sardinians) but that looks unlikely on first sight (why would Finns and Iberians have the same level of gender-bias unless it's approx. zero?)"

CEU-like sex bias all around Europe becomes plausible if we assume that most of Europe did not experience any gender-biased major demographic events, beyond the pan-European expansion of groups with steppe ancestry. This leaves room for various independent population histories around the continent, they just lacked sex-biased migrations.

That Sardinian paper is the same I linked, and they determine the eastern sex-biased ancestry of Sardinians not based on Y-DNA or mtDNA but based on autosomal DNA and X-chromosome with its different inheritance pattern. Sardinian Y-DNA is actually quite WHG-EEF (I2, R1b-V88 which more likely has EEF links than steppe links, G2), certainly more so than CEU's, yet the WHG-EEF shift is in the X-chromosome and not the autosome.

Matt said...

@Maju, as far as I'm concerned you're welcome to link / repost / include. Just to explain a bit more background though so I'm not causing confusion -

Like I believe Alberto's table your referencing, it's based on PCA data by Davidski run through the nMonte R script written by Ger Huijbregts.

In this instance, the PCA data it's run on is Davidski's "Days of High Adventure" PCA combined with the Globe10 in one datasheet.

The "Days of High Adventure" is a PCA which IRC this was one Davidski built around the ancient West Eurasian dna primarily itself - http://eurogenes.blogspot.co.uk/2016/08/ancient-vs-modern-day-west-eurasian.html where the Globe10 is based on global data including the ancient West Eurasian dna and modern.

(So more of DoHA's dimensions should be relevant to structure between the ancient samples where Globe10 is more dominated by the differences between world populations and fewer dimensions will relate strongly to the ancient West Eurasians.)

Main Post for this PCA was: http://eurogenes.blogspot.com.au/2016/11/days-of-high-adventure.html and the World Population Data for "Days of High Adventure" is here in the comments: http://eurogenes.blogspot.com.au/2016/11/days-of-high-adventure.html?showComment=1479031871735#c5088487017669989883

I previously tried to use the "Days of High Adventure" PCA alone with nMonte and it was if anything more specific to the subregions of Europe - http://eurogenes.blogspot.co.uk/2017/02/first-look-at-baltic-and-ukrainian.html?showComment=04517454865405705885#c4221879531969042837

I would say though I find the results in my table hard to believe literally as actually true unless there somehow really was a totally unsampled essentially CHG wave into Europe through Greece and the Balkans at around the same time as the Yamnaya culture and we've somehow totally missed it until now (which seems really unlikely from what I know).

(Seems likely there must have been some regional survival of HG in Europe, as when we look at the median haplotype donation in Cassidy et al, with the only high coverage ancient HG, Loschbour, measuring direct ancestry then that peaked in British Isles populations - (http://www.pnas.org/content/suppl/2015/12/23/1518445113.DCSupplemental/pnas.1518445113.sapp.pdf page 60) not the Lithuanians and Eastern European population who seem to have the most HG affinity by f3 stats or IBS or Basque who are high relative to MA1 affinity. But it is seems hard to imagine that the figures I found from nMonte with literally 3 way combinations of local hg+early neolithic+CHG will be what is real, given all the work around EHG / MA1 shift.)

For more about nMonte with the "Days of High Adventure PCA" I also found it was pretty regionally specific with the Middle Neolithic data I found http://eurogenes.blogspot.co.uk/2017/01/qpadm-tour-of-europe-mesolithic-to.html?showComment=04517454865405705885#c4947706076116291674 and it seems to tend to reproduce Bronze Age population averages as being broadly closest to present day people from a region - http://eurogenes.blogspot.co.uk/2017/01/qpadm-tour-of-europe-mesolithic-to.html?showComment=04517454865405705885#c6898693074086301965. So it's interesting why it is seems so specific.

Davidski said...

I would say though I find the results in my table hard to believe literally as actually true unless there somehow really was a totally unsampled essentially CHG wave into Europe through Greece and the Balkans at around the same time as the Yamnaya culture and we've somehow totally missed it until now (which seems really unlikely from what I know).

I expect we'll be hearing a lot about this hypothetical CHG or even Neolithic Iranian wave deep into Northern Europe over the next few weeks, as it's the only hope left for many including Maju, until the next big paper finally kills the idea.

Maju said...

@Matt: thank you very much, I'll try to include all possible credits. I won't probably write anything until later in the week anyhow, my back aches too much with this rainy weather and also have some stuff to do.

"it seems to tend to reproduce Bronze Age population averages as being broadly closest to present day people from a region"...

That's interesting because I also have the impression that the "demographic stabilization" seems to happen across the board around the Bronze Age (in some places earlier but nowhere I can detect at later times, maybe Romania??) Of course there should be some minor changes afterward but nothing big (and I attribute that to the "aristocratization" of conquest since Metal, which does not look anymore for lands to work themselves but for servants to work for the lords: ethno-linguistic change still happens but it isn't caused anymore by demographic change but rather by elite domination).

Maju said...

@Saikorth: are we then on the same page re. your X-DNA argument? That there is no apparent gender bias and that therefore we should expect (roughly) mtDNA pool ≡ Y-DNA pool ≡ nDNA pool? If so, doesn't this imply that we should expect to find high R1b-S116/U106 where mtDNA includes high, modern-like, frequencies of H (and other haplogroups), which for the Neolithic should be basically the area of modern France (with penetration in the Basque Country, maybe also the lower Rhine, Denmark, etc., and excluding the Mediterranean coasts)?

"That Sardinian paper is the same I linked"...

My bad again. :(

I should read more carefully, really.

Maju said...

@David: "I expect we'll be hearing a lot about this hypothetical CHG or even Neolithic Iranian wave deep into Northern Europe over the next few weeks, as it's the only hope left for many including Maju, until the next big paper finally kills the idea."

It has nothing to do with anything else I may be arguing for or against, it's just an odd detail that caught my attention. Latvia is too remote, has too low densities and is too peripheral to matter and anyhow how would the CHG/IranNeol thingy matter at all? We do know for a fact that the "teal" component (aka Caucasus-Baloch, Gedrosian, Caucasus-Central Asia, Highland West Asia...) is an important marker in Kurgan peoples and their influence since Yamna, whether there is something like that in Latvia Neolithic or not seems almost totally irrelevant for the wider picture. So please do not over-read in what I said, I just find it curious and also find it curious the rest of the affinities that Alberto and Matt seem to have detected.

But of them, the most interesting one for me is the SHG element, surprisingly strong, not the CHG one, much less in the remote and pretty much irrelevant ancient Neolithic Latvian population.

Davidski said...

The rest of the affinities that Alberto and Matt seem to have detected.

They didn't detect anything.

There was a migration of a steppe population very similar to Yamnaya to Latvia during the Late Neolithic/Bronze Age.

This was argued well enough in the Jones et al. paper using Latvia LN1 as the relevant ancient sample.

I backed up those findings in Jones et al. with a method that they didn't try: modeling based on formal stats. It's all here.

http://eurogenes.blogspot.com.au/2017/02/no-game-changer.html

What is most certainly not there is any sign of an CHG or Iran Neolithic migration to Latvia at any stage.

Maju said...

@David: so what you're saying is that the 43% CHG in Latvia-LN1 (Alberto's sheet) is because of Kurgans. Fair enough, but how do you explain that they are 41% SHG and only 16% EHG? Aren't you reaching your conclusion of "interchangeability" with Yamna on an analysis that lacks SHG (and hence produces aggrandized EHG frequencies)? Has anybody tried even comparing with Yamna directly?

Also what about the 14% CHG in Latvia MN1 and the 12% CHG in Ukraine HG1 (and 9% in Ukraine N1)? While there is a clear increase of CHG in Latvia with CW, a small fraction seems to have been there earlier, even since Paleolithic times in the case of Ukraine.

In any case, I insist, my main interest was all the time in the surprisingly high SHG frequencies in all the samples (except Latvia MN, which are low or even zero in SHG). Why do we get 55-78% SHG in Latvia HGs and 66% SHG in Ukraine N1 (much larger than in Ukraine HG1)? I'm not arguing for anything, just making questions about the complexity and affinities of pre-Kurgan Eastern Europe, which are less straightforward than expected (would be high EHG everywhere but nope).

[PS- again getting issues with the Captcha: if you commit an error, in this case link tag was not properly closed, you don't are "doomed"]

Davidski said...

@Maju

Fair enough, but how do you explain that they are 41% SHG and only 16% EHG? Aren't you reaching your conclusion of "interchangeability" with Yamna on an analysis that lacks SHG (and hence produces aggrandized EHG frequencies)? Has anybody tried even comparing with Yamna directly?

Also what about the 14% CHG in Latvia MN1 and the 12% CHG in Ukraine HG1 (and 9% in Ukraine N1)? While there is a clear increase of CHG in Latvia with CW, a small fraction seems to have been there earlier, even since Paleolithic times in the case of Ukraine.


This is very easy to explain: it's 90% bullshit.

Maju said...

That's what your algorithm, in independent hands, seems to say. "It's 90% bullshit" is the very definition of shitty explanation.

Davidski said...

It's OK to experiment and get things wrong. It's not OK to believe anything that just happens to land online.

Fact is, you're out of your depth. I could spend hours trying to explain to you how to interpret various results, but you wouldn't understand any of it.

At this point what you need to do is to give up trying to understand anything, and just take my advice, which, not by coincidence, is exactly the same as what the people who sequenced Latvia LN1 communicated in their paper.

Unless you do this, you're going to make a total fool out of yourself, and discredit yourself as a useful source in this area of study. That's because this year alone, paper after paper will come out arguing, for good reason, that Yamnaya-like people moved en masse from the steppe into the East Baltic and Poland, with no CHG or Iran Neolithic farmers in sight.

Maju said...

I'll ignore all that of "making a fool of myself", I'd rather risk that than become a statue: we only live once. But:

1. Khvalynsk/Yamna has a c. 50% Caucasus component that was absent in Volga EHG previously. I believe that even you must agree that it is evidence of Zagros Neolithic (or similar) genetic influence in the area, right?

2. AFAIK Khvalynsk/Yamna have not any detectable SHG (and tell me if I'm wrong in this, please) but it's all EHG. So this is contradictory with your claim of Latvia LN (high SHG, low EHG) being exchangeable by raw Yamna, at the very least on first sight.

3. The Caucasus element seems to have permeated to some other Eastern European populations, prior to Kurgan expansion, even if at quite lower frequencies than in the Volga. That's an interesting bit on itself, although I guess independent confirmation would be needed to make sure it is something real and not just a "ghost".

Rob said...

Maju & Dave

From the paper
"The latest Neolithic sample in our Baltic time series, Latvia_LN1 (5,039–4,626 cal BP), which was found in a crouched burial of the type associated with the Late Neolithic Corded Ware culture [21], falls near other Late Neolithic and Bronze Age European and Steppe samples in PCA analysis (Figure 2A). In ADMIXTURE analysis, it is composed of the blue component (Figure 2B), which is predominant in all of the older Latvian samples, but also a green component, which is maximized in hunter-gatherer samples from the Caucasus."

Davidski said...

You're taking that quote out of context if you think it argues for a migration of CHG people to the East Baltic.

The point they are making there is that Latvia LN1 is not native to Latvia, but arrived there from an area where CHG admixture was present, or at least a lot more pronounced, like the steppe.

The paper clearly argues that there were population movements of Yamnaya-like people from the steppe to the East Baltic, just as other recent papers have argued that there were population movements of Yamnaya-like people from the steppe to Central Europe.

Maju said...

But their HG component is not Yamna-like, i.e. EHG, but, per the quote and also in the tests done in this discussion and its parent one, it seems it is mostly SHG. And that's a bit of a problem for your "explanation", David, in which Yamna and Latvia LN are "interchangeable". They may be comparable but they are clearly different in their HG component.

I think that your problem is that this FACT adds a layer of complexity you are uncomfortable with, because you like explanations simple in certain very specific ways: if SHG was in pre-Kurgan Eastern Europe and also shows up in "Kurgan" Latvia, you may have to think further, admit a greater layer of complexity than the ultra-simple "triangular model", which just doesn't work well enough.

What's your answer to all this issue? "Bullshit" (sic). OK, it seems every inconvenient piece of data is "bullshit" but what's that kind of reaction? Like Donald Trump tweeting before breakfast. Have you thought running for Queen of Australia? I think it's the right time.

(BTW, unsubscribing, I'm not having this "bullshit" anymore).

Grey said...

one way to square the circle might be

1) steppe *horse* dude expansion in the *swampy* Baltic was more elite conquest than full displacement cos terrain - hence SHG survival

2) the expansion linked one end, the Baltic, with the other end, the Caucasus

3) SHG chief gets a chg wife as a result of this linkage and has an shg/chg kid

(doesn't have to be a chief but given the distance it seems more likely - was it a rich burial?)


#

separately it would be cool if someone cleverer than me could figure out given conditions of:
- equal bride exchange over a border over multiple generations
- the population on one side of the border is smaller than the other
- over time brides on both sides will have ancestry from the other side
then what is the equilibrium mtdna percentage for the smaller population for particular starting population ratios

that could provide a "natural" baseline which would might make the rest easier to believe

capra internetensis said...

Here's some really simple math:

Yamnaya samples n=15
Autosomal: ~half EHG
Mitochondrial: ~half EHG (7/15)
Y-chromosomal: massive recent founder effect

Significance of evidence for sex bias: nil
Quantifiability: lol

Davidski said...

@Maju

The HG component in Yamnaya and Latvia LN1 might be somewhat different, because Latvia LN1 is likely from a different, more western part of the steppe, so its HG input is more WHG-like.

But the point is that Latvia LN1 is from the steppe, and your evidence that it isn't from the steppe is not valid. Not only does it contradict Occam's Razor in a major way (where is that CHG wave deep into Europe?), but it's based on methods that are likely to produce spurious results at such a fine scale level, due to overfitting, differences in sample quality, or several other reasons.

Davidski said...

@Capra

It's likely that some of the mtDNA lineages that you're counting as EHG actually arrived on the steppe from the North Caucasus with CHG-rich women.

You're largely focusing on the deep Holocene or older affinities of these haplogroups, but Yamnaya only formed ~3,500 BC. That's a problem and your mistake.

Thus far, there is absolutely no evidence of any male migration from the Caucasus/Near East onto the steppe at that time. All we have are clues that women from the North Caucasus were incorporated into steppe societies.

These clues come from mtDNA, isotopes and archeology. So unless you can actually come up with evidence that the Caucasus brides theory is false, or at least that it doesn't sufficiently explain the formation of Yamnaya, that's the best theory we have, and it actually makes sense on several levels.

capra internetensis said...

@Davidski

Sauce for the goose is sauce for the gander.

'Farmer' lineages can be EHG origin, but then so can 'EHG' lineages be of CHG origin. U4a and U5a might be absorbed by CHG, but then so might R1a or R1b.

Likely enough there was sex-biased gene flow between them - there is skeletal evidence for this in both directions at different sites. But the present evidence can fit perfectly well with plenty of "Caucasian husbands" as well as "Caucasian wives". And if you want to claim otherwise, show us your math.

Davidski said...

But the present evidence can fit perfectly well with plenty of "Caucasian husbands" as well as "Caucasian wives".

No it can't. Kurgan lines from across space and time during the Bronze Age look fully European: R1a, R1b and I2a2.

It's the mtDNA that shows southern admixture. The simple math you want is misleading for the reason I outline above.

At best you're looking at a few Caucasian husbands here and there. As far as I'm aware though, no one has actually found them yet. But they have found the Caucasian wives with isotopic data.

Davidski said...

And sure, some of that R1a, R1b and even I2a2 in the Kurgans could in theory be from the Caucasus, just like some of the U5 and U4 might be from the Caucasus.

If so, however, how did only European-specific Y-HGs make it onto the steppe at that time? Where's the J2 for instance?

Wouldn't you expect at least a couple J2, maybe G, or some other clearly southern Y-HG? Why is the southern admixture clearly evident only in the mtDNA?

Y-DNA founder effects don't really explain too well why only Eastern European-specific Y-HGs migrated from the Caucasus onto the steppes.

capra internetensis said...

None of the known EHG lineages, including 4 different ones at Samara alone, occurs in Yamnaya either. How much male CHG component there may have been originally is unknown. We are going to need more ancient DNA before we know exactly what happened on the steppe.

The isotope data I recall showed that women who had recently been previously eating a marine diet were found at a some interior sites during Catacomb and Yamnaya times. Not women from the Caucasus during the formative period of the steppe pastoralist population.

The skeletal evidence I recall was for the EEF rather than Caucasian people, e.g at one Cernavoda I cemetery there were mainly Proto-Europoid men with Mediterranean women, at an Usatovo cemetery the other way around. Also the Igren Sredny-Stog cemetery had Proto-Europoid women and a mix of Mediterranean and Proto-Europoid men (in the latter case perhaps they could have been Caucasian).

But I don't have data for the Maykop-Novosvobodnaya and earlier cultures north of the Caucasus, which are the most relevant ones.

Davidski said...

We are going to need more ancient DNA before we know exactly what happened on the steppe.

So what? Female exogamy from farmer groups to steppe herders is a legitimate hypothesis and a very good one until then.

We have no Caucasus-specific Y-HGs on the Bronze Age steppe, but fresh Caucasus-specific mtDNA, and examples of exactly such female exogamy at other sites where steppe-derived people settled.

http://eurogenes.blogspot.com.au/2015/05/high-female-mobility-in-bronze-age.html

So I'm not sure why you're about to pop a testicle by arguing against this legitimate and common sense hypothesis? Seems like there's something hiding under your objectivity, most likely the preference for another outcome.

capra internetensis said...

Nope. I'm a North American of British ancestry and I don't give two shits about how many of my male ancestors in Eastern Europe 5000 years ago came from which side of the Caucasus. Prehistory is interesting but extremely remote from anything I actually care about.

Bad reasoning annoys me. Whatever, I'm done now.

Davidski said...

You're wrong though. It's not bad reasoning.

The dribs and drabs of data we have from the steppe generally fit with the expected behavior of steppe people.

So the female exogamy hypothesis has a very good chance of being confirmed, like it was in the Lach Valley in Bavaria with the steppe-derived Bronze Age folks there.

Rob said...

So far the Majkop mtdna doesn't quite look like Yamnaya , but it's a small dataset
Western Yamnaya and catacomb look very EuroHG

Davidski said...

I've got a spreadsheet here with Catacomb and Yamnaya mtDNA.

http://eurogenes.blogspot.com.au/2016/10/steppe-boys-farmer-girls.html

jv said...

This the image of the Yamnaya "chief" from Samara Russia Kutuluk River. He was yDNA R1b1a2a2(Z2103) and mtDNA H6a1b. http://i61.tinypic.com/1zcpteb.png

jv said...

Davidski, thanks for the graph. I'm curious about the 4 mtDNA H6 Catacomb Culture in Ukraine. Is that the full sequence? That is a lot of mtDNA H6 in ancient Ukraine! Yamnaya & Poltavka & Srubnaya Cultures in Russia has 3 H6's. Siberia has 2 H6, Okunev & Andronovo Cultures.......looking forward to the Bell Beaker DNA results and wonder if a mtDNA H6 will show up in them. H6 is associated with Corded Ware but mixed Beaker/CW marriages are a possibility.