search this blog

Friday, January 30, 2015

Half of our ancestry comes from the Pontic-Caspian steppe


Here's the latest teaser for the new David Reich et al. paper on the ethnogenesis of present-day Europeans. It's part of an abstract for a seminar to be held by Professor Reich at Jesus College, Oxford, on February 9. Interestingly, it argues that migrations from the steppe resulted in a ~50% population turnover across northern Europe from the late Neolithic onwards, which is very much in agreement with recent discussions on the topic at Eurogenes (for instance, see here).

By ~6,000-5,000 years ago, a resurgence of hunter-gatherer ancestry had occurred throughout much of Europe, but in Russia, the Yamnaya steppe herders of this time were descended not only from the preceding eastern European hunter-gatherers, but also from a population of Near Eastern ancestry. Western and Eastern Europe came into contact ~4,500 years ago, as the Late Neolithic Corded Ware people from Germany traced ~3/4 of their ancestry to the Yamnaya, documenting a massive migration into the heartland of Europe from its eastern periphery. This steppe ancestry persisted in all sampled central Europeans until at least ~3,000 years ago, and comprises about half the ancestry of today’s northern Europeans. These results support the theory of a steppe origin of at least some of the Indo-European languages of Europe, and show the power of genome-wide ancient DNA studies to document human migrations.

Source: Ancient DNA documents three ancestral populations for present-­day Europeans


Update 11/02/2015: Massive migration from the steppe is a source for Indo-European languages in Europe (Haak et al. 2015 preprint).


Haak et al., Massive migration from the steppe is a source for Indo-European languages in Europe, bioRxiv, Posted February 10, 2015, doi: https://dx.doi.org/10.1101/013433

646 comments:

«Oldest   ‹Older   401 – 600 of 646   Newer›   Newest»
Sonic Reducer said...

Are there any plans for a similar study of Gumelnita- Karanovo (so) cultures? Possible homeland of Y I1 and MT V ?

Marnie said...

@Shaikorth:

That's very interesting. Will have to think about that before I comment.

Thanks.

Marnie said...

@Sonic Reducer:

regarding Gumelnita- Karanovo (so) cultures, you might find this paper interesting:


Radiocarbon Dating of the Neolithic Lakeside Settlement of Dispilio, Kastoria, Northern Greece

Yorgos Facorellis, Marina Sofronidou, Giorgos Hourmouziadis

https://journals.uair.arizona.edu/index.php/radiocarbon/article/view/17456

it's open access.

Shaikorth said...

It has to be noted that the specific breakdown of European affinities for those ancient genomes in MyOrigins is interesting, since many of them (like Loschbour and Anzick-1) clearly predate formation of modern European groups and should share affinities with all of them through WHG and ANE. I just pointed this issue out in Razib's blog, he's involved with MyOrigins and may get more out of the results.

WesternPonticSteppe said...

Hi, you removed this article: http://bga101.blogspot.ro/2012/07/pinpointing-loci-specific-genomic.html but I am very interested in my chromopaint to check the Siberian/East Asian chunks. Do you have a link to the image archive? Thanks

Matt said...

Davidski: It sounds like the ANE-rich population that mixed with the EHG to create the Yamnaya was already mixed itself, and probably had more ANE than present-day Armenians.

Sounds possible, although that would make me wonder why it was described as population like Armenians, and not Lezgins or even Abkhazians or Chechens. Perhaps only name recognition (everyone knows of the Armenians)? We will know soon.

Unknown said...

Marnie,
Micro-blades date back 50kya, with AMH dispersement. The north Eurasian link with Europe is probably found in pressure flaking.

http://www.academia.edu/458820/Microlithization_in_Eurasia_a_Brief_Review_on_the_Microblade_Reduction_Technology_and_its_Significance_as_a_Behavioral_Threshold_of_the_Modern_Humans

http://popular-archaeology.com/issue/june-2013/article/stone-tools-in-india-have-implications-for-early-dispersal-of-modern-humans-out-of-africa-say-researchers

Unknown said...

I think it could be described as Armenian and EHG if it were that they were less ENF and more ANE than Armenians, but replaced more than 50% of the aDNA. That would push back to Armenian /EHG, at close to 50/50.

Davidski said...

Oaie,

Sorry, I can't find those files. But you had two years to check your results.

Unknown said...

Ryu
"probably quite a significant fraction, of ANE is going to penetrate from EHG and SHG in the far NE."

A possibility I've always maintained needs to be explored before making simple, blind conclusions

If true ; then the "yamnaya intrusion " scenario would be little more than an episode of localised gene flow; undoubtedly one which also went west to east ; given the possibility that at least some of the Neolithic type admixture found in yamnaya could be from SEE.

pnuadha said...

@MIke Thomas

I think I found those papers...

Awesome, mind sharing them? The one I posted was about two or three years ago. I also something similar at the ASHG conference which would obviously be more recent.

BTW whats your account for ANE in north-western Europeans ?

Well, just based on my r1b theory I expect that Yamnaya to Hungary followed by Hungarian BB to Northwest Europe, to account for most of the ANE.

I think some ANE came by way of the CW and later migrations; maybe there was even a small amount of ANE in scotland due to mesolithic migration from Scandinavia. But because Northwest European ANE is so similar to Northern Central Europe, and Norway, there must have been some quick moving ANE carrying people that rushed past Central Europe. The way to rush through North Central Europe is by yamnaya up the danube then further West under the name of eastern bell beakers.

A more interesting question is the distribution of ANE in Eastern Europe post Neolithic and how CW had its ANE. I don't think yamnaya are at all necessary...

Unknown said...

Hhhmm I agree with the possibilities . I furthermore entirely agree with the centrality of the Carpathian basin (Hungary) as a "womb for Europe " (at least northern europe) something repeating throughout history .

Wrt to the papers , the ones I found are rather older - Keyser (in discussing the Papuans) and Jorge

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC379223/


http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1288178/


Maybe u can recall the more recent ones ?

Kurti said...

@ryukendo
"
By the way, the latest estimates place other populations at >66% Ymanaya with high-WHG Yamnaya, up to ~85% Yamnaya with the low-WHG estimates, too... such as the non-ENA portion of Kazakhs.
"

English is my fourth language so I often don't get the exact meaning of some sentence. Tell me that you said the non ENA portion of modern Kazakhs would come close to Yamna figures?

Kurti said...

If what I understood is also what you said, than my estimation of Andronovo beeing similar to the West Eurasian portion in Uzbeks, Kazakhs etc is probably correct?

Kurti said...

@ryukendo

"I think lower WHG and higher ANE would drive them east, while the ENF % has not changed, so their NS position is the same."

That was exactly my thoughts from the beginning. This is why I thought the WHG would be less and ANE more so they would shift towards east while ENF staying constant on the typical figures for Balkanians, North Caucasians or Ukrainians.

Chad said...

Kurti,
If Yakuts, according to the recent paper, represent the ancestral Turks, then that makes a good starting point. Assuming that the Yakut get a bit of their ANE, all their ENF, and maybe all their WHG from IE expansion to Siberia. I would make the Yakut about 12% ANE and 88% EEur, prior to contact. That would make the West Eurasian part of Kazakhs at about 42%ENF, 25%WHG, 33%ANE. This portion would come out to about 36% of their ancestry, with the remaining 64% the ANE/EEur mix. It does look very much derived from the west and not the south. By way of the fact that there was probably already an ENF/ANE mix, with minor WHG already present. Factoring that in, it certainly looks like our Yamnaya mixed with those folks, then Turks came in.

Before anyone get's all worked up, remember it's all hypothetical and speculation. Just a best guess. If you want to pick it apart, use your own math, not just NO....

Chad said...

By the way.. looking at Mongols, it's almost the exact same percentage of ENF to WHG. These figures probably aren't far from the truth.

Chad said...

There is too much WHG in the West Eurasian part. If Andronovo came from the south, thereby removing Yamnaya from the equation, then you're going to need way more WHG in the natives. I don't see any hunters being 40-50%WHG in Kazakhstan.

postneo said...

http://www.annualreviews.org/doi/full/10.1146/annurev-linguist-030514-124812

Primarily a linguistic argument with some perfunctory geographical/archeological info thrown in for forms sake.

Its devoid of archeological migration routes from the steppes to Anatolia, Iran and south asia ! All action happens in Europe.

For Indo-Iranian they hide behind Renfrew.

For Anatolia, they just draw a line from the Pontic steppe to Romania and leave the crucial entry to Anatolia completely to imagination.

"4200–4000 bce, coinciding with the Suvorovo–to–Cernavoda I migration and the sudden end of the tell cultures of Old Europe;"

leaving language aside, the above route could be of relevance for R1b in Europe.

Unknown said...

Cernavoda is a mix of Pontic kurgan and varna related folks. There connections back and forth with Anatolia, covered by a couple papers I've seen. One of the later groups, Ezero, does show connections to both the Carpathians and Troy.

Unknown said...

Their*

Kurti said...

"West Eurasian part of Kazakhs at about 42%ENF, 25%WHG, 33%ANE."

Nice that you admit my estimation of 25/45/30 might be true.

So do you understand now why I "wrestled" the already "professionally" calculated propotions?

Davidski said...

Nonsense.

Kazakhs aren't just Yamnaya/Turkic. They have recent ancestry from the Near East and South Central Asia associated with the Turkic and Islamic expansions. This is obvious in their uniparental markers.

Kristiina said...

When I used the word microblade culture I had in mind specifically Butovo culture which flourished west of the Urals c. 10 700 – 8200 BC when N1c did not exist.
This article is interesting http://www.phil.uni-greifswald.de/fileadmin/mediapool/histin/Ur-_und_Fruehgeschichte/Projekte/Russland/Hartz__Terberger__Zhilin_2010.pdf

They say that ”in a broader perspective the production of microblades can be traced back more than thousand years further east. In the southern Transbaikal area microblade production is proposed for the site Studenoye 2 and Ust’Menza 2 dated up to c. 17 000 BP. In further regions of Siberia and in Japan the production of regular microblades and the use of pressure technology can be identified on Upper Palaeolithic sites dated after the Last Glacial maximum and before start of Greenland Interstadial 1. In general there is no doubt that specific microblade production was in use in eastern Eurasia earlier than further west and also spread over to North America. But this does not necessarily mean that this technology was taken over in western Eurasia from the east.”

Microblade techniques spread also to Scandinavia and Germany as they say that ”in the western Baltic there has been a long debate as to the introduction of regular microblades and pressure technique with special focus on the introduction of handle cores. Handle cores are a specific type of elongated microblade core with a narrow front. A re-evaluation of the evidence by Olofsson argues for the start of an elaborated microblade technology and slotted bone tools in the younger Maglemose Culture in southern Scandinavia. The same holds true for northern Germany where on sites such as Loop 1 and Lammershagen 10 an advanced macroblade production developed in the late Boreal parallel to a microblade technique with tiny regular microblades stuck from single platform conical cores.”

Kurti said...

@David

You don't have any evidence to back this up. I go that far and claim the Balkans and Iberia had more of influc during the Islamic period than Kazakhstan!

Having "unipaternal" typical Near Eastern Haplogroups is your "prove" for that? Come on man. Come up with something more convincing.

Davidski said...

Don't worry Kurti, the Yamnaya genomes are coming soon.

Matt said...

Thinking (probably over thinking) on possible variation in Yamnaya ancestry in Northern Europe to arrive at the average of 50% -

I would qualify this by saying that as ANE does not differ that much between Northern Europeans - 14% in SE England, 17% Poland, 18% Lithuania, taking the Eurogenes estimates. So assuming that all the ANE is via Yamnaya as an axiom (unless there's an important non-Y source of ANE), it is hard to have very disjoint contributions from Yamnaya between them.

To qualify what I mean by "that much", say Yamnaya had 35% ANE (which I think is high but reasonable to accomodate the other balance of WHG and ENF ancestry). A 40% Y, would then be predicted to have 14% ANE, so far so good, but 75% Y population would then have 26% ANE, which is pretty high when we look at the highest remaining Europeans. If Yamnaya instead had 24% ANE (which would seem low), then 40% Y would become 9.6% ANE, 75% Y would be 18% ANE.

It seems like it would make more sense to have North Europeans vary from around 0.45 Yamnaya, 0.55 Middle Neolithic in SE England to 0.6 Yamnaya, 0.4 Middle Neolithic in Lithuania, based purely off the assumption that all ANE is via Yamnaya. But then 50% depends on who is binned into Northern Europe... If there was even some geneflow, even 10% from BHG / SHG, then levels would have to fit even more even between (assuming no other flows of ANE into Northwest Europe), and no geneflow from BHG / SHG doesn't seem as likely as it did (it looks like Middle Neolithic populations may have had more WHG than Yamnaya now (say if MN is like Gokhem 50 WHG and 50 ENF, while Yamnaya is say 30 ANE, 30 WHG, 40 ENF) so Yamnaya ancestry alone seems unlikely to explain how WHG rich the Baltic is).

Kristiina said...

I think that ANE started to spread to the west already via the Butovo related cultures. I have heard somebody say that Reich et al would have analyzed Mesolithic Karelians. It would be very interesting to know what their WHG and ANE ratio was. I think that the highest concentration was in Finland and Karelia and diminished further west in Baltics, Scandinavia and Germany. Do we know how much ANE Motala has?

Davidski said...

It's not likely that all of the WHG around the Baltic today is entirely of Yamnya origin, but a good chunk of it probably is, and almost all of the ENF might be too.

The situation in Northwestern and Central Europe is probably somewhat different, with most of the ENF being of local Neolithic origin, but most of the WHG coming from the east after the Neolithic.

Simon_W said...

I've thought about it. Imho the strongest evidence in favour of a northern /steppe origin of PIE is its relationship with Uralic. However, afaik a genetic relationship between the two language families has not been proven. If the observed similarities between IE and Uralic are mostly the effect of strong early IE influence upon early Uralic, then a West Asian origin of PIE is still possible. It would be impossible if there were clear traces of Uralic influence in, say, Hittite. But to my knowledge there is no such indubitably Uralic trace. There are just similarities between IE and Uralic which may have originated in either of the families.

A major problem for the theory of a steppe origin of PIE is: The steppe foragers have adopted the West Asian way of life as herders. Whose language was more likely to prevail? Of course the language of the dominant cultural model. Especially if the West Asians made up 50% of the population. I can't see how the hunter-gatherer language should have prevailed under such circumstances. So the West Asian highlands are still the more likely source of PIE.

I've also been thinking about the Corded people. There is the problem that they were typically long faced and high skulled, which is the opposite of the typically broad faced and low skulled Yamnaya people. I can't see how the Corded people could have been 3/4 Yamnaya descended. But there is an alternative. Reich doesn't claim that the Corded people were indubitably Yamnaya descended. He rather holds that they were either Yamnaya descended or from a closely related group. I think the latter is more likely to be the case. Foragers in the Baltic were typically long faced. Foragers in the northern Pontic steppe used to be very broad faced. Hence the ancestors of the Corded people were probably closer related to Baltic hunter-gatherers. They were ANE rich EHG in either case. But to make them genetically more similar to the Yamnaya people we also need a strong Armenian-like admixture. This needn't be a problem. If the Yamnaya people even as far north as Samara were 50% Armenian-like, then the steppe groups in the very south, Kemi Oba and Lower Mikhailovka, were probably even more West Asian. And from these an influence may have diffused northward to give rise to the Corded people. So, interpreted that way, Corded Ware and Yamnaya were descended from similar sources but not one from the other.

Somehow we have to make sense of the archeological findings, too, we can't simply ignore them, and I think my above suggestion may open up a way for doing so. If the custom of Kurgan burials originated in Transcaucasia and spread to the steppe together with the Armenian-like influence, then we're not forced to assume that the Corded Ware „Kurgans“ spread together with the Corded people and the Corded Ware material culture. And that is exactly what the archeological record suggests. In fact, to me it looks as though the earliest beginning of the Kurgan burial custom in central Europe was in the Baalberge group of the TRB. These had occasionally erected Kurgan graves. And interestingly, one of the typical mtDNA haplogroups of the late Neolithic / EBA, haplogroup T1, was also found in the Baalberge culture. This is hardly a coincidence! They didn't have the EHG haplogroup U2, but they had Kurgans and T1. From this root the Corded Ware burial rite slowly developped. And quite independently from this process, the strongly EHG descended Corded people spread together with the Corded Ware material culture a couple of centuries later.

Davidski said...

Motala12 has a maximum of 19% ANE. But that's still way too low for it to be a candidate, because its WHG ratio is far too high. The math simply doesn't add up.

The people who contributed the vast amount of ANE to the modern European gene pool had to have an ANE/WHG ratio of something like 40/60. That's what Yamnaya or CWC genomes probably have.

Simon_W said...

What the evidence from ancient DNA seems to suggest, is a movement of ANE rich EHG descended groups westwards on the one hand, with the Corded people and presumably also with the Globular Amphora culture. These were West Asian admixed and probably spoke an IE language. On the other hand, BR1 and BR2 in the Carpathian Basin were probably descended from an earlier move of EHG southwards along the Black Sea. Because especially BR1 had hardly any West Asian admixture, so she probably wasn't descended from West Asian admixed Yamnaya, but from earlier, purer steppe groups. The expansion of the Bronze Age from Southeastern Europe washed them up into the Carpathian Basin. But if BR1 spoke IE is unclear to me. It also has to be said that her ANE was probably weaker than in central European Corded people. This earliest ANE wave into the Carpathian Basin wasn't as strong as the ANE wave into central Europe. Only later there followed other steppe groups with more ANE and with West Asian admixture.

As for R1b, it looks to me as if that expanded from West Asia onto the PC steppe, but this doesn't necessarily mean that the main branch of European R1b was descended from the steppe. R1b may easily have spread into several different directions and ended up in southwestern Europe at quite an early date.

Davidski said...

Simon,

There's no evidence that 100% Near Eastern migrants made it to the Samara Valley.

What's being reported is that a partly Near Eastern, ANE-rich population mixed with the local hunter-gatherers there to produce the Yamnaya nomads by 3000 BC.

Unknown said...

David
"NE/WHG ratio of something like 40/60. That's what Yamnaya or CWC genomes probably have."

So are you now suggesting that he "native " precursors Id CWC kight already have had this ane? (Without a need for marked intrusions from yamnaya or further east)?

Davidski said...

Yamnaya > CWC

It says so in the abstract I posted above. This conclusion is probably based not only on overall genetic structure alien to Neolithic Central Europe, but also Y-DNA sequences and mtDNA genomes.

What this means is that CWC got their ANE from Yamnaya. It also means that at some point you'll have to abandon your theory that CWC sprang from elements native to Central Europe.

Balaji said...

@Matt,

You were wondering why Reich et al. described the ANE-rich near-Eastern ancestors of the Yanmaya as Armenian-like instead of Lezgin-like. I think it is because they think that the Yamnaya acquired their language from their Near-Eastern ancestors and not from their EHG ancestors. Lezgins and other North Caucasus people do not speak Indo-European languages but Armenians do. Therefore Reich et al. did not describe the near-Eastern ancestors of the Yamnaya as Lezgin-like though they probably had ANE levels more similar to Lezgins than to Armenians.

Animal domestication and pastoralism is a development of agricultural societies not of hunter-gatherers. The Yamnaya would have learned pastoralism, metal working as well as their language from their more technologically advanced ancestors. Simon W. also pointed this out in his comments above.

It appears that most people here believe that Indo-European languages originated in the European steppe and spread from there to Western Europe and to Asia. Reich and Patterson, however, believe that Indo-European originated in the Caucasus and that the Maikop passed it on to the Yamnaya. Nirjhar believes that PIE is from the Zagros area of Iran. Simon W. believes that PIE originated in West Asia. I believe that PIE is from the Indian Subcontinent.

Davidski said...

David Anthony was listed as one of the co-authors on the SMBE 2014 abstract that first broke the news about the Yamnaya genomes from the Samara Valley. So he's likely to be involved in this new paper that we're all waiting for.

Meantime, he's just put out this paper on the PIE question...

http://www.annualreviews.org/doi/full/10.1146/annurev-linguist-030514-124812

Unknown said...

David
"CWC sprang from"

CWC did not spring from anything . It was a wholly new epiphenomenon which was regionally varied; and not a homogeneous block. There can be no doubt it has a complex mosaic of elements which can be considered, both, native/ continuous and new/ "alien". Yet I was previously arguing that many of the pastoralist changes seen in the period actual were incipient during the late neolithic (especially in areas like the late Cucuteni) and not wholly new transplantations from the steppe.

What it might have meant to be in a corded ware community in one village in Germany probably meant something different to one ifurther East, and indeed different yet even to a close Neighbour

Unknown said...

Question is ;
A) was there indeed a *large scale * movement from the steppe ( as there can be no doubt of smaller scale admixtures )
B) is there evidence for movement from central - eastern europe; as well as "southern" regions, to the steppe, in return.


Hopefully reichs and future papers who'll certainly clarify

Davidski said...

Bullshit.

There was a detailed study done a few years ago on Corded Ware, and the conclusion was that the cultural links were so strong between the different Corded Ware communities that they all had to have sprung from the same source. Kristian Kristiansen talks about it in this vid.

http://www.youtube.com/watch?v=-C4XsA5ovbg&list=PLAXoDomeFLX90fTHi0W8lYBtEoZHSBH2i&index=7

This is now backed up by ancient DNA.

Unknown said...

Kurti,
That's for Andronovo, not Yamnaya. I'm standing by Yamnaya at 39enf/35whg/26ane. You can't have Yamnaya at 30% ANE and keep Corded Ware in Northcentral Europe.

Unknown said...

Simon,
Baalberg is NOT Kurgan. Those are barrows. They're used by several megalith groups. Even Britain has them by 4000 BCE. There's no ANE in those groups.

Ryukendo K said...



@ Matt

I have already done some new estimates, under different scenarios of what 'N.Europe' means. I will repost them here:

I strongly suspect the level of Yamnaya ancestry was estimated by the relative level of ANE. The biggest unknown at this stage is what '~50% in Northern Europe' means. The 50% Armenian thing is still true as Nick said, and any estimate with ~38% ENF and >7% ANE will fit this criterion.

1) If northern europe is conventional, aka scand + NE Europe, these areas have approx 16-18% ANE on average. Doubling this results in 32-36% ANE for Yamnaya. With ~38% ENF, result in <30% WHG.

2) If Northern Europe incl. Germany and British isles, the figures don't change much, dial ANE down to 32% and WHG up.

3) If Northern Europe refers to everywhere north of Pyrenees and Alps, this incl pops like Bulgarians and French which are approx 12% ANE. Say we take 14% as the average of all. Then we get 28% ANE, ~38% ENF, 34% WHG.

So we have a range.

To add to the above, a figure of >36% ANE would place the most ANE Euros at 50% Yamnaya, a figure of 30% Would place them at 60%, and approx 27% ANE at 65%. I expect that would be the likely range for ANE figures. Since ENF is gonna be fixed at ~38% for the 50% Armenian rule, just vary WHG accordingly.

ANE will probably not score at 27% or lower, because that would make many Europeans ~75% Yanaya, and if many Euros are ~100% CW they would have said this, instead of making discontinuity between CW and N.Euro an apparent point in their abstract.

Lastly, I think the authors would be much less sure about the ~50% figure in the main text than the abstract would suggest, because people like finnish have close to the max amt of ANE found in Europe, despite having <7% R1a and other indications of IE ancestry, which should indicate that using ANE to gauge in an addition-subtraction way is not a good idea in all cases, esp for the pops around the finnish. I expect them to hedge their bets by pulling a Mike Thomas somewhere in the text, leaving open the possibility of other scenarios.


@ Kurti

Taking WHG as a gauge of Yamnaya ancestry in the non-ENA proportion of Kazakhs, for Yamnaya with 36-32% ANE, thus 26-30% WHG, that portion in Kazakhs would be 68-85% Yamnaya, as the non-ENA portion of Kazakh is ~21% WHG.

For lower ANE at 30-27% and higher WHG at 32-35% WHG, the non-ENA portion turn out 66-60% Yamnaya.

I'm doing mental calcs, typing quickly, hope my arithmetic is on point.

@ Kristiina @ Shaikorth @ Marnie

Wow a treasure trove! Thanks for taking the time to reply.

I'm busy now, but have a lot of thoughts, will reply later.

Davidski said...

rk,

Finns are irrelevant, because they're Uralics with ~5% post-MA-1 Siberian ancestry, so they won't fit any of these models, like they didn't fit the ANE/WHG/EEF model.

At some point you'll have to acknowledge this and quit bringing them up.

Mike,

I'm yet to hear a coherent argument from you on this topic.

Unknown said...

I've told u my argument . But for you its disappointingly agnostic - even nihilistic ; and rather dull. No invading hordes of nomads

Ryukendo K said...

@ Davidski

You know as well as I do that that is not a relevant issue at all, especially when you can get Kazakhs and Kalash, who have a great deal more ENA, to fit in your 3-ratio triangle.

There are no formal stats for ENA vs ANE, but that has not stopped you from trying to get a handle, even if it might stop the scientists. And you appear to have succeeded, and we believe you. I don't think it should stop the rest of us, who've followed you at each step of your reasoning, and are just extending this logic further?

The fact that finns have 95% of the ANE of Balts while having massively less uniparental from Yamnaya is still an issue. The fact that Balts have more ANE than poles despite have less uniparental from Yamnaya just demonstrates the point further.

Well, this is not going to affect the estimates the scientists make, which is still gonna be ~60%, because formal stats can get the proportion but not the source. But people reserve the right to leave a bit of open-mindedness what those ANE proportions actually mean. In my case, I think SHG and BHG mean the actual figure is somewhat lower, in practice.

Davidski said...

Yeah, I read something funny from Frachetti once.

He was arguing that R1a might have ended up in the Andronovo kurgans through genetic drift from South Asia, but not the genetic drift as actually defined in genetics, but a different kind of genetic drift, as in it drifted up there from the south. Hilarious shit. :)

Davidski said...

rk,

Unless Finns received their language from a 100% ENA Siberian population, which seems very unlikely, then they are significantly more Uralic than ~5%, and Uralics shared a border with Yamnaya.

So there's a very good explanation why Finns don't have much R1a or Yamnaya-like mtDNA, and yet could appear mostly Yamnaya in some models.

You can't say the same about Balts or even Estonians, because they come from former Corded Ware strongholds, mostly influenced by migrations from other Corded Ware strongholds, and carry a lot of R1a.

Lithuanians and Estonians will top the f3 shared drift stats with the Corded Ware and Yamnaya genomes, mark my words.

Kristiina said...

"Unless Finns received their language from a 100% ENA Siberian population, which seems very unlikely,"

I think that I heard a leak that ancient Karelians were more WHG than ANE, so 100% ENA cannot be true and even if they were 100% ANE I suppose that they would not be 100% ENA. I think that areas from which N1c came out were close to Samara proportions.

postneo said...

@simon_w

On Uralic. Are there uralic loans in european daughter languages? Can they be resolved as early loans?

If not then we have a symmetric case of uralic never being borrowed for any of the daughters in europe or asia.

If they exist then an anatolian/south caspian/baluch source of PIE is likely.

Ryukendo K said...

@ Kristiina
Where did you hear it from? The buzz surrounding this has become an echo chamber, with all its assoc distortions.

@ Daviddski
For the balts I don't disagree. That wasn't the point I was making.

The closest Uralics to the Finns are 25% East Asian and 22% ANE. Finns are 5% East Asian, which is a fifth dilution, so it makes perfect sense to suggest that the dilution of Uralic ancestry into E.Euro down to finns cannot account for more than a fifth of the ANE in Finns. 80% of their genome cannot be fitted as Uralic, so the ANE has to come from somewhere. Since it doesn't make sense to suggest that 80% of the finnic genome came in intrusively from balts, it makes sense to say its local.

Ryukendo K said...


Time to wade into deeper water. Sorry if I lose some people.

@ Shaikorth

Thanks for your comments. They make a lot of sense.

To add on, I'm actually trying to get at a highly specific issue: does IBD and IBS influence ADMIXTURE estimates apart from, and independent of, their effects on allele sharing? To understand what I mean, I'm specifically addressing the fact that the ADMIXTURE algorithm relies on allele frequencies at loci for analysis WITHOUT resort to any kind of information as to where the loci are, or whether or not they are contiguous in a segment of any kind. Thus this kind of info should not be available to ADMIXTURE.

Therefore, ADMIXTURE should be insensitive to time esp. time as it relates to decay of segments. Instead, it should be sensitive to patterns of loci where populations tend to share similar allele freqs at same loci, a.k.a collinear drift. Therefore, ADMIXTURE should be sensitive to time only as much as time changes the importance of drift in recent pops. To get how this is relevant, consider the ANE K=8 which is prob David's most impressive accomplishment. Pops like Ket, Kalash and Kabardin are not gonna be high-ranking in IBD/IBS by a long shot, neither are they gonna have sig IBD/IBS wih Mal'ta, esp compared to e.g. czechs and poles. And significantly, ADMIXTURE identifies local drift among Siberians, Europeans, other subcontinental pops as the most sig dimensions of drift in the dataset, thus producing the familiar NE Euro, SC Asian, etc. clusters, which makes perfect sense considering that these dimensions structure the dataset the most due to recent collinear drift. But when ADMIXTURE is forced to consider that dimension where Mal'ta differs from everyone else, aka differences of Mal'ta's allele feqs will everyone else's allele freqs, ADMIXTURE has no problem inferring the needed cluster despite the presence of highly drifted, closely-related subsets of pops such as SC Asians and NE Euros in some dimension orthogonal to Mal'ta.

Nirjhar007 said...

@Postneo
''On Uralic. Are there uralic loans in european daughter languages? Can they be resolved as early loans?''
Its Only Balto-Slavic which seems to been suggested to have Uralic loans and of course they have clear Uralic substratum influence in their Phonology.

Ryukendo K said...

@ Shaikorth

Sorry, its a very complicated and subtle point, I'll finally get to the end which relates to how to interpret ancient genomes with modern ADMIXTURE clusters.

To continue, this makes perfect sense of the fact that in unsupervised runs, IBS, aka old shared ancestry within or between a pop, results in the tendency to produce a component at lower Ks, while IBD which is genealogically relevant is gonna alter the proportion of shared alleles so much for a few people its gonna produce a huge dimension of variation for a small cluster of ppl that only appears at high K due to its accounting for very little of the structure in the dataset. Therefore, do you agree that IBD and IBS affect components only and only due to their effects on allele freqs?

The fact that ADMIXTURE hates linkage disequilibrium might relate tangentially to the point here about it trying to do as little as possible with the position of alleles, though figuring out how this works mathematically is making my head hurt.

Therefore, say an ancient genome is unearthed, say Mal'ta. Its so old that IBD and IBS with modern pops is a comparatively meaningless concept. The clusters ADMIXTURE assigns it to in an unsupervised run should have more to do with collinear drift than anything else, shouldn't it?

Suppose we have another ancient genome that is closer in time, say Iron Age. This is enough for the score in IBD and IBS to mean something, and also for it to be assigned very modern-looking ANE, WHG and ENF. However, say it scores a lot of North Sea, despite being from Siberia. Surely this means that, in that dimension where British and other pops high in North Sea differ from evveryone else, this genome also differs? The IBD and IBS for the ancient genome are prob going to localise to pops with the same ratios of WHG, ANE and ENF, but that has very little to do with anything. A N.Cauc genome would share highest IBD/IBS with pops from the middle east and SC Asia, but that has everything to do with similar ratios/generalized ancestry across the ancestral genome, and little to do with actual collinear drift, no?

Whew, that was tiring.

Ryukendo K said...

@ Shaikorth
To summarise, in the same way that ADMIXTURE can overlook/be insensitive to time issues and IBD/IBS and use collinearr drift in the direction of Mal'ta to produce estimates of ancestry from Mal'ta in the presence of other clusters with collinear drift for population groups at the same 'taxonomic level', aka WHG and ENF, ADMIXTURE primarily uses collinear drift to target even later ancient genomes and divvy up their ancestry in the presence of unsupervised clusters, no?

Ryukendo K said...

@ Kristiina

Thanks for your comments.

Could you explain to me one thing? I never understood why MRCAs for Y-haps in only a subpopulation, like E.Euros, made any sense, because it makes sense that the common ancestors for most haplogroups in a subpop should lie completely outside of that region/population altogether. E.g. R1a's MRCA in India would not be found in India, R1b's MRCA in US populations would be much older than the time US populations have actually existed, and would actually be in Europe, etc.

It seems to me like the distr of N in N.Europe might testify to multiple migrations from a source outside where the MRCAs of N in Europe would be found. Or am I totally wrong?

Lastly, this pt to Shaikorth too, I would like to push back somewhat against the idea that E.Asian existed in Europe since early times, because the pop density of WHG is 10X lower than Karitiana and all genomes from spain to Hungary are incredibly homogenous. If there is any discernible 'exotic' element it would have leached all over. If ENA was there it would have leached into EHG, and this would be highly noticeable in D-stats such that I think the authors would have mentioned it.

Ryukendo K said...

@ Davidski
Generally, liberals sound especially airheaded when they deride pop gen as 'its all social constructs!'. But here
We are actually approaching the boundary where 'closeness' has become a very fuzzy term and involves a lot of statistics. IBD, IBS, PCA of genotype data, PCA of ADMIXTURE results, neighbour-joining, etc, are probably gonna produce slightly different results for different reasons.

Balts are gonna score very high in f-stats, but in PCA of genotype data they're not gonna be the closest, its prob gonna be poles and ukrainians. And Ymanaya's gonna be very far from everyone generally.

Personally, I would prefer a sum of squared-differences statistic between Yamnaya's ADMIXTURE stats and modern pops' ADMIXTURE stats, and are Balts gonna be the closest in such a scenario? Almost definitely no.

(For the rest, to give a basic overview what I mean is subtract the ADMIXTURE results of Yamnaya from every other pop, and square so that you get rid of negatives and proportionize the distances, and then sum, which gives a pythagorean distance. This is gonna be better because we have some handle on where the ENF, ANE and WHG comes from, as opposed to just the proportions alone, e.g. is the ENF from the near east of Med? Is the ANE local or from Siberia?--as ADMIXTURE is sensitive to this to a certain degree.)

Ryukendo K said...

@ Marnie
Well, we all have our bad days.

Once again, I would appreciate your input on the admittedly very technical questions.

Shaikorth said...

rk,

I'm not suggesting there was ENA in WHG and I'm not at all certain if it was in Mesolithic Europe, the morphology and mtDNA of EHG's could be just ANE-related if populations with mainly ANE and ones with mainly ENA looked alike. However there were clearly areas in Europe where there was ANE and areas with only WHG, so boundaries existed. And if there was an ANE zone, why not an ENA zone too? Originally the idea of ENA in EHG was proposed in the comment section of this blog, because the talks about the upcoming Reich paper didn't give breakdowns for EHG's WHG and ANE (and Yamnaya's for that matter, it became 50/50 "Karelian/Armenian"), and such difficulties they have previously attributed to ENA presence.

EHG's through Yamnaya would spread it all over Europe and that would explain why ADMIXTURE runs with a less common population set can show things like mainly ENA modal component of Hazaras being higher in CEU than in Brahui (0% in Sardinia -> 10% in CEU -> 25% in Kargopol Russians who don't get an unique eastern component, making the ENA in Europe a quantitative issue in this run). Note that these results would differ for CEU and Orcadians if there were Lithuanias in that run because they'd probably form a cluster absorbing "Hazara" and only Kargopol Russians would be more eastern than it.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4055572/bin/pgen.1004393.s012.xlsx


ENA in EHG would also explain those Dai-related f-admixture signals that pop up all around Europe compared to Basques, who do have some ANE, in all likelihood more than Dai do. Anyway, even if there was ENA in EHG's it is clear that they aren't the only source of it in modern Europe because the highly ENA-related ADMIXTURE clusters that may appear in Europeans vary a lot (and don't necessarily stabilize in language families) as do Ld-based estimates for ENA ancestry (Lazaridis' ALDER test, Yunusbayev 2014 Turkic migration paper).

Re. Balts and shared drift with steppe folks, regardless of their actual Yamnaya ancestry they'll peak the f-stat due to their homogenousness, just like they'd get more shared drift with Chuvash than Kargopol Russians do. The big question is whether Ukrainians and Belorussians get higher f3-stats than Czechs.

Kristiina said...

Ryukendo,

”The closest Uralics to the Finns are 25% East Asian and 22% ANE.”
Which group do you mean? I did not expect Mordovians to have 25% ENA. I know that Samis have more Siberian than Finns but I did not expect them to have that much. In any case, Saamis and Mordovian are linguistically the closest groups to Finnic groups (Estonians, Livonians, Vepsans, etc.)

”80% of [Finnish] genome cannot be fitted as Uralic”
What you mean with this Uralic genome? How can you know what it is?

”It seems to me like the distr of N in N.Europe might testify to multiple migrations from a source outside where the MRCAs of N in Europe would be found. Or am I totally wrong?”

You should actually ask Shaikorth and not me, but my comment is that the only age estimations for Finnish N1c that I remember are here http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2986641/, but I do not know how up to date this paper is, and, in any case, their estimations should be too high. There are also other unexpected things such as their age estimation for Finnish R1a1 which is older than their age estimation for Finnish N1c.

Yes, I think it is difficult to say for how long a certain haplotype has been in a certain area, and we know that people can quite easily change place. So, there are surely many uncertainties. However, usually you do not claim that a haplogroup or its subgroup comes from an area where the variance is low. Unfortunately, I have not seen a detailed breakdown of variance of N1c, but in any case the variance should be lowest in the Arctic where the frequency can be very high. This seems to apply also to European R1b.

”Lastly, this pt to Shaikorth too, I would like to push back somewhat against the idea that E. Asian existed in Europe”
My idea is that it is specifically the northeastern microblade cultures that spread both ANE/ ENA to the west. We know that the distance between east and west shrinks the more to the north you go, and these post glacial cultures I am talking about reached East Europe only 10 000 BC. Their population numbers were surely very small, so it would be just natural that their contribution to the modern populations is extremely diluted and concentrated in the north (Northern Scandinavia, Finland and Northern Russia). With this, I do not claim that yDNA N did not spread any ENA, for example my Southeast Asian ancestry could very well be due to yDNA N.

Unknown said...

Highlighting nonlinear patterns in population genetics data sets....

http://www.nature.com/srep/2015/150130/srep08140/full/srep08140.html

Shaikorth said...


”It seems to me like the distr of N in N.Europe might testify to multiple migrations from a source outside where the MRCAs of N in Europe would be found. Or am I totally wrong?”

You're not. The rare N-P189.2 is distant from N1c lines phylogenetically and seems to appear mainly in populations with no other N, so at least one separate migration happened. If we speak of Northern Europe we speak about just N1c1, and it may have come all at once and split west of Urals in which case it might be Mesolithic or early Neolithic in the eastern fringe of Europe only, OR subclades came separately in which case it's very recent everywhere it appears in large numbers (Finnic, Baltic) since MRCA's of those clades are recent.

Even if the MRCA of all European N1c1 clades was outside Europe it means that its presence in the Baltic region likely postdates Yamnaya and CWC (possibly by quite some time). OTOH if that MRCA of all present European N1c1 lived west of Urals it becomes likely that the main paternal lineage of Yakuts is a bottlenecked remnant of a post-Mesolithic migration from Europe and that there was also a migration towards Bhutan, because a N1c1 lineage (N-Y9022) that split earlier than theirs can be found only west of Urals. That is probably too extreme to seriously contemplate without ancient DNA evidence.

Kristiina, re. Y-DNA datings of R, N or generally anything take a look at Molgen or other places where SNP-based datings from full sequences are discussed. The datings from old studies based on few STR's can be safely forgotten if there are better alternatives around.

Unknown said...

David

"rk,

Unless Finns received their language from a 100% ENA Siberian population, which seems very unlikely, then they are significantly more Uralic than ~5%, and Uralics shared a border with Yamnaya.

So there's a very good explanation why Finns don't have much R1a or Yamnaya-like mtDNA, and yet could appear mostly Yamnaya in some models."

Sounds like special pleading instead of facing the elephant in the room ?

Unknown said...

PostNeo


"On Uralic. Are there uralic loans in european daughter languages? Can they be resolved as early loans?

If not then we have a symmetric case of uralic never being borrowed for any of the daughters in europe or asia.

If they exist then an anatolian/south caspian/baluch source of PIE is likely."

The earliest undisputableoand are indo Iranian to Uralic ; although Anthony would have us believ these loans were as early as early Indo european itself

But loans tell us little definite information. It's svject to chance ; dynamics and doesn't take heed of the fact that other intermediaries might be long extinct

Davidski said...

Which elephant in the room is that Mike, the one you've been avoiding since the start here; that almost all of Europe was affected by massive migrations after the late Neolithic?

The only difference between Finns and the rest is that Finland was affected by somewhat different eastern migrants, speaking Uralic languages and carrying ENA as well as ANE.

Which part of that do you still not understand? Would you like a Youtube video of a puppet show to help you grasp the basics here? I've got some spare socks I can turn into Yamnana nomads, Uralics, Neolithic farmers, etc.

Matt said...

@ RK: I have already done some new estimates, under different scenarios of what 'N.Europe' means. I will repost them here

I appreciate this, however I am interested here in what this is saying about the Middle Neolithic Europeans, given then range of likely Yamnaya contributions to European populations and the likelihood of SHG admixture in Lithuanians and Balts, not in the constituents of Yamnaya, which are we can estimate really no better or worse than we previously could. It may be that most of the WHG in many Northwest Europeans comes from MNE not Yamnaya or Mesolithic Baltic survivors (like the Comb Ware culture who seem likely to have contributed to Finns and probably Balts, Finns looking plausibly like Balts who were then culturally dominated by Saami like groups after this synthesis took place, cf http://www.unz.com/gnxp/on-the-limes-of-the-modes-of-production/).

Unknown said...

David
Ha ha . At least you know you have a career as a joker if/ when your theories become unstuck

Marnie said...

@Chad

Good find. (Also, thanks for the microliths summary paper you put up yesterday.)

I've been suspecting that PCA has limitations with "nonlinear" effects for a while. Nice to see this discussed in a paper.

your link:
http://www.nature.com/srep/2015/150130/srep08140/full/srep08140.html

from the paper:

"In practice, the PCA approach has been shown to be very powerful and reliable, although it suffers from two major drawbacks:

"i) the curse of dimensionality, i.e., the problem of finding information in datasets characterised by an overwhelming number of features over samples, which is a typical problem in population genetics;

"ii) difficulties associated with revealing nonlinear patterns hidden in a high-dimensional space.

"As a consequence of these problems, PCA is occasionally unable to detect differences between groups of individuals, even with prior knowledge that such differences exist. There is also the case in which although we do not have such prior knowledge, dissimilarities characterised by some unknown nonlinear feature relationship may be present in the high-dimensional space, but because PCA is unable to detect them, they cannot be identified."

Thanks for posting this important paper.

Chad said...

Matt,
I believe that there are remaining groups outside of the easily recognizable MNE groups. There are still hunter groups roaming around at the time of Corded Ware. It is possible that these have way more WHG than ENF. If all of the U5 in those forager sites, then Corded and Beaker are telling us anything, it could be that they liked to mix with those outside of dense areas as much as they did with actual MNE like Baalberge, Salzmunde, etc. I think that my Oracle is showing that as well. My oracle is 50% Yamnaya, 25% Gok2, and 25% Loscbour. There had to be more WHG floating around than a simple Gok2-like stuff.

If our Yamnaya is correct, then someone like me, at 46%WHG, has an extra 6% over Gok2. That is also clear on the PCA's where Gok2 can only cover WHG to about the 40% mark, if we are half Yamnaya. Everything else becomes additional input from western groups that lack ANE.

As for the Baltic/Comb Ceramic stuff, I am pretty much in agreement that it probably on contributed to the Baltic area, with SHG also contributing up north, with minor introduction to the south and west during Germanic and Viking movements.

Anyway, someone really needs to test those Blatterhohle and Ostorf hunters.

David,
Do you remember which Gok farmer was supposed to be more WHG than Gok2? I cant remember, but I think it was Gok7. I scored higher with them, than Gok2, on the ancient sample Oracle.

Davidski said...

Not really relevant to the current discussion though, as the K8 with which I use to estimate ANE proportions isn't a PCA analysis, and the discovery of the massive shift in European genetic structure after the Neolithic didn't come from PCA analyses.

Davidski said...

Gok7 only has 10,000 SNPs available. Not enough.

Marnie said...

Has anyone tested the idea that Gok2, Iceman, other Alpine ancient DNA samples, weren't farmers, but alpine hunter gatherers? Or maybe alpine pasturalists (who were originally part alpine hunter gatherers)?

Marnie said...

@Davidski

"the discovery of the massive shift in European genetic structure after the Neolithic didn't come from PCA analyses."

Which method did you use?

Chad said...

Okay. That's too bad. I'm not inferring that the K8 suffers from linear equation related problems. I am just showing the short comings to basing everything on lines and numbers from here to there. I try to avoid that myself. We need ancient genomes that cover a wide space and time to put it together, which the K8 is a leap forward to. As more samples become available, I think that you'll be able to refine it to a point it is as good as anything anyone else uses. The paper we are waiting on will certainly help us to create models of populations. The only issue is migration period stuff that mixed it up a little more.

Chad said...

Marnie,
They have Basal Eurasian. That isn't found in any Mesolithic sample.

Marnie said...

@Chad

"They have Basal Eurasian."

So what? What if "Basal Eurasian" is just a Pan-Mediterranean-Near East population that ended up "going Alpine" during the Neolithic?

That's what the Iceman looks like to me.

Krefter said...

Chad,

You're ancestry is mostly from northwestern Europe which is very far away from Russia. We don't know what your bronze age eastern Euro ancestors picked up on their long journey to west Europe.

French can fit as 50/50 Yamna/Gok2 nicely but obviously are not.

The spread of Yamna-ancestry in Europe is not the result of unrelated admixture events. It was gradually carried from east-west and kept gaining native baggage on the way.

Euros sharing Yamna ancestry shared probably also means Euros share other forms of ancestry which was added along the east-wet Journey(BHG, EEF, etc.).

My point is if Lithuanians have BHG, there's a good chance Irish and Portuguese do to.

Hopefully you understand what I'm saying. It's hard to articulate.

Marnie said...

@Davidski

Sardinians have almost no "ANE", yet their ydna hgs are about 20% R1b.

So how can the R1b hg be correlated with "ANE"?

Shaikorth said...

Matt, the most recent dating of Baltic-Finnic languages in Finland and Estonia is so late (little more than 2000 years) that it's unlikely the border regions of Corded Ware that reached Finland were absorbed by Uralic speakers. Some idea of what they were in absence of ancient DNA may perhaps be found in lactase persistence frequency. Saamis show little of it, even less than Volga-Ural populations like Udmurts, but for Finns it is very high, close to British and Scandinavian level. East Finns have lower levels than West Finns but even they have considerably more than North Russians and even Estonians (who are more Corded Ware-related).

Regarding genetic markers and languages, if either the Karelian (EHG) or Kola Peninsula Oleni Ostrov (3500bp, over a millennia before there's evidence of Saami languages in Kola) have ENA, it's pre-Uralic in far Northeast Europe and not a good marker of a present language's spread. N1c1 is probably going to end up a decent Uralic marker regionally though.

Marnie said...

@Shaikorth

Lastase persistence is a good example of a highly non-linear trait (in time.) It has shown to be "up selected" over very short time scales.

I don't think we can understand very much about Finns from lactase persistence alleles.

Anonymous said...

@Chad Rolfson

It's Gok4, IIRC.

http://www.nature.com/ncomms/2014/141021/ncomms6257/images/ncomms6257-f2.jpg

Shaikorth said...

Marnie, true that LP is not much, but in absence of ancient (autosomal and Y-)DNA such speculation based on modern population traits is really what we have.

Marnie said...

@Shaikorth

"Marnie, true that LP is not much, but in absence of ancient (autosomal and Y-)DNA such speculation based on modern population traits is really what we have."

"is really what we have" ?

which is not much. No data. For the Finns/Balts question, anyway.

Unknown said...

Krefter,
Baltic ancestry is probably about nil, in Iberia. It should be pretty small in all of Western Europe.

Marnie said...

@Shaikorth

Why not run the ncNCE software on Finns and other Baltic pops instead?

I had a look at the results they were getting for Japanese populations. It looks very promising.

Unknown said...

Marnie,
The same way someone can have R1a or b in Africa, with about zero ANE. A couple of guys had successful lines with a loss of aDNA. R1a and b split after ANE comes into play. If one is linked to ANE, the other one was too, at least in the beginning.

Davidski said...

Not all Sardinians lack ANE.

Those from the HGDP that I've analyzed have virtually 0% ANE (the 4% that Lazaridis et al. gives them is an overestimate, due noise created by using a single ancient genome as a reference IMO). However, other Sardinians are more similar to south Italians, and probably do carry small amounts of ANE.

Matt said...

Shaikorth: Some idea of what they were in absence of ancient DNA may perhaps be found in lactase persistence frequency.

Sorry if I'm not following, your thinking is that there was no fusion of Corded Ware and Comb Ware in Baltic Europe followed by change to Uralic languages in Finns (with some, not much dna change in Finns) instigated by migration because the Finns have high lactase persistance and have adapted biologically to drink milk?

Krefter said...

@Chad
"Baltic ancestry is probably about nil, in Iberia. It should be pretty small in all of Western Europe."

Iberia probably got most of its ANE from Northwest Euro-type pops. So, whatever northwest Euros have Iberian have at a smaller amount.

Northwest Euros may have gotten most of their ANE from Baltic-types, so whatever Balts have northwest Euros have at a smaller amount(don't forget about Iberians).

It seems you think Yamna-type ancestry largely spread independently in each region of Europe, and I dis agree. How do you explain Norse and Iron age Brits(also modern Irish) clustering right next to Polish and Belorussians in PCAs?

Is it because of independent events? Or is it because of common bronze age central-east Euro ancestry? NW Euro's Yamna ancestry probably didn't only go through the Balkans, because how do you explain excess WHG without hunter gatherer admixture? Hunter gatherers were living up north ~4,000-5,000YBP, not in the Balkans.

I don't know though I haven't thought about it much. It's a very complex situation.

Marnie said...

@Chad

"The same way someone can have R1a or b in Africa, with about zero ANE. A couple of guys had successful lines with a loss of aDNA. R1a and b split after ANE comes into play. If one is linked to ANE, the other one was too, at least in the beginning."

Or, the alternative . . . that ANE is correlated with R1a (weakly), but not really correlated with R1b.

"If one is linked to ANE, the other one was too, at least in the beginning"

Doubt it. R1b - R1a split was probably at least 15 thousand years ago.

How come Q and R rich pops, which split about 30,000 years ago, have almost no shared components on admixture, but you are assuming that R1a and R1b would necessary both be strongly correlated with a particular "component" ("ANE" in this case) ??

Marnie said...

@Davidski

"Not all Sardinians lack ANE."

"Those from the HGDP that I've analyzed have virtually 0% ANE (the 4% that Lazaridis et al. gives them is an overestimate, due noise created by using a single ancient genome as a reference IMO)."

Come on. If you can somehow get from 4% ANE in Sardinians and suggest that that explains the 20% r1b hgs in Sardinians, then you can also assume only very weak correlation between any of the autosomal components you have, and hgs.

House of cards.

Krefter said...

Marnie please be humble when you're newer to this than everyone here? Everyone has to admit there's a possibility what they're saying is wrong. Even I can see many of your statements are a result of being uninformed.

Y DNA P itself is very correlated to ANE. Sardinian R1b is almost entirely typical Italian R1b-U152. My guess is that Roman and later Italian men were elites-rulers in Sardinia and got more women.

Admixture based on modern pops is differnt from admixture based on ancient genomes. Y DNA R and Q pops are full of ANE, and MA-1 aka Mr. ANE had R and lived 24,000YBP.

Marnie said...

@Chad

"Baltic ancestry is probably about nil, in Iberia"

See these papers regarding the connection between Iberia, Dordogne, Northwest Europe, and the Baltic:

Miller, Rebecca, "Mapping the Expansion of the Northwest Magdalenian, Quaternary International, 30 May, 2012

Valentin, Boris et al, Magdalenian and Azilian Lithic Production in the Paris Basin: Disappearance of a Programmed Economy

Sönke Hartz, Thomas Terberger, Mikhail Zhilin, "New AMS-dates for the Upper Volga Mesolithic and the origin of microblade technology in Europe", Quartar 57 (2010): 155-169

Ballin, Torben Bjarke, Alan Saville, Richard Tipping and Tam Ward, "An Upper Palaeolithic Flint and Chert Assemblage from Howburn Farm", South Lanarkshire, Scotland: First Results, Oxford Journal of Archaeology, Volume 29, Issue 4, pages 323-360, November 2010.

The Repopulation of Northern Europe, book chapter, Felix Riede, in
The Oxford Handbook of the Archaeology and Anthropology of Hunter-Gatherers, Edited by Vicki Cummings, Peter Jordan, and Marek Zvelebil




Marnie said...

"Marnie please be humble when you're newer to this than everyone here?"

I'm not being particular un "humble".

Just asking some [rather obvious] questions.

Also, Krefter, maybe my style is a little Silicon Valley like, not European like.

There has to be some room to ask hard questions. I'm sorry if you're interpreting my style is being too direct.

Marnie said...

"Marnie please be humble when you're newer to this than everyone here?"

I'm not being particularly un "humble".

Just asking some [rather obvious] questions.

Also, Krefter, maybe my style is a little Silicon Valley like, not European like.

There has to be some room to ask hard questions. I'm sorry if you're interpreting my style is being too direct.

Chad said...

Marnie,
I'm talking about Baltic hunters with ANE. This was probably after 9000BCE. Anyone that went from the Baltic to Western Europe, was pre-ANE. There is no ANE in Western Europe before the Copper Age.

Davidski said...

Marnie, there's also R1a in Sardinia, including R1a-Z280 lineages that look like they came straight from Russia.

But of course they didn't. They were introduced from southern Europe by people who were usually typical southern Europeans in terms of genome-wide genetic structure, with low levels of ANE.

This is why it's possible for Sardinians to carry ~20% of Y-haplogroup R and barely register any ANE.

You can't compare this situation to Copper Age continental central and northern Europe, which was affected by direct and massive, for those times, gene flow from the steppe.

Davidski said...

Hang on, I thought people in Silicon Valley were supposed to be smart?

Chad said...

Marnie,
SNPs say that R1b and R1a split between 10-14kya, and it sounds like the new paper puts the split at 10kya. They are from Central Asia, and certainly ANE. What else would be in Central Asia between 10-30kya? Look at Kurds.. they have a good chunk of I2 y-dna, but almost no WHG. It isn't detectable at the EEF level of Basal Eurasian.

Marnie said...

@Davidski

I think it's one thing to be direct in asking questions about technical points.

It's quite another to be constantly insulting the intelligence, sanity and person of women who try to participate on this blog.

You're indirectly accepting public money for your research and in treating the women on your blog differently, you're breaking the law. At least in the US and Canada.

Chad, all I can say is that you better have a look at the small print of what it means when you accept public funds for research in the US and Canada.

Anyway, Davidski, you're apparently obfuscating the questions that I asked you:

What method are you using to set up the "dramatic turnover of the population of Europe by "ANE" after the Neolithic?

If you're so willing to allow for a non-linear effect in the upscaling of R1b hgs in Sardinians, to explain they're lack of ANE even with 20% R1b hgs, then why wouldn't this non-linear effect appear all over Europe. And if so, why are you using linear algebra methods to analyze population movements in Europe?

Davidski said...

Here you go Marnie, to get you started...

http://eurogenes.blogspot.com.au/2014/11/3-pop-analysis-featuring-19-ancient.html

http://eurogenes.blogspot.com.au/2014/12/ane-is-primary-cause-of-west-to-east.html


Unknown said...

Marnie
Can you spell out to a nube (me) what exactly yr non-linear method is ?
Although a science background; I humbly lack the extent of expertise in PCA and other statistical analyses some others here have ; so have been mostly taking their calculations Prima facie; but only critiquing their interpretation of the data

Romulus said...

Kurds do not have a "chunk" of Y DNA I, there are two samples that have been done on Kurdish Y DNA, one is from a total sample of 95 North Iraq Kurds that found 16 individuals of Y DNA I, the other is from a sample of 95 Kurdish Jews that found 0 Y DNA I. The study did not differentiate between subclades of I.

Chad said...

Marnie,
I don't accept public funds. I am not who you think. My name is exactly who I am.

Marnie said...

@Chad

"SNPs say that R1b and R1a split between 10-14kya, and it sounds like the new paper puts the split at 10kya. They are from Central Asia, and certainly ANE. What else would be in Central Asia between 10-30kya? "

OK. No dispute there.

I'm pointing that many R1b rich populations do not have much "ANE". For example, Armenians, Tuscans, Sardinians and Basques.

That's all.

So it's a problem to assume that R1b rich populations brought the "ANE" component to Western Europe.

There could be some substructure in other R1b associated autosomal components that have caused the "Eastward Shift" in Europeans since the Mesolithic.

Chad said...

Romulus, it just proves my point that someone can have double digits of a haplogroup and pretty much lack the Mesolithic component that was associated with it. Here is a list at a Kurdish site.. There's more out there. I'm just not investing more time into it. Someone else can look it up, out of their own curiosity.

Haplogroup I
4x I-M170 (Iraqi Kurds in Stenersen et al., 2004; based on Athey's Haplogroup predictor)

9x I-M170 (Zaza from Turkey in Nasidze et al., 2005)

14x I-M170 (Kurmanji from Turkey in Nasidze et al., 2005)

1x I-M170 (Iranian Kurds in Malyarchuk et al., 2013)

1x I2-M438 (Iranian Kurds in
Grugni et al., 2012)

1x I2a2a-M223 (Kurdish village Dogukoy*/Central Anatolia in Gokcumen et al., 2011)

1x I2a2a* (old I2b1*; Z161+, L1228-, L1229-, L1230-, L1226-, L699-, L701-, L702-, L703-, L704-, M379-)(Sorani from Sulaymaniyah/Iraq)

1x I2a2b-L38 (Kurdish village Dogukoy*/Central Anatolia in Gokcumen et al., 2011)

Chad said...

Marnie,

"I'm pointing that many R1b rich populations do not have much "ANE". For example, Armenians, Tuscans, Sardinians and Basques"

Armenians have 15.5% ANE, that is not a good example for you. Southern Europe was no doubt more densely populated and didn't get the same Copper/Bronze Age impact. As for the Basques, look how far they are from the ANE source population. Plus, R1b probably wasn't in Sardinia or Iberia prior to 2000BCE. It would've been diluted down before getting that far. I think the high density population in France had something to do with it. That is where you see the ANE dip. The Alps and the southern half of France.

I could cherry pick out the West Germans at 15%ANE, and Irish at 15%, or English at 14%. I think Irish are the best example. They had the largest population boom in the Copper Age and Early Bronze Age. Their population increased 4-fold, and is probably why they have some of the highest ANE in Western Europe. Every Irish result I've seen falls between 15-17.2% ANE. The people with the highest R1b in Western Europe, are about the highest in ANE.

No one else could've brought all that ANE to Britain, except a group of people led by R1b males.

Marnie said...

@Mike Thomas

It's usually easier to define what a linear system is than a non-linear system, so I'll start there.

A linear system is one that can be analyzed with methods of linear algebra.

http://en.wikipedia.org/wiki/Linear_algebra

So, for example, in this discussion, R1b is ANE correlated, and if there is 20% R1b hgs in Sardinians, and equal numbers of founder men and women, you would expect there to be *at least* 10% ANE in Sardinians (accounting for the possibility that there was no ANE contribution from mtDNA founders):

expected percent ANE in Sardinians = 0.2(u) + 0(v) = 0.2*(0.5) + 0*(0.5) = 0.1

a=20% r1b men carrying ANE

b=0% women carrying ANE

u=0.5 and v=0.5, population proportions between men and women, assuming they are even.

if u=0.2 and v=0.8 (for example) more non ANE carrying women pass on their autosomes than ANE/R1b carrying men

then the expected percent ANE in Sardinians = 0.2*(0.2) + 0*(0.8) = 0.04

or 4% . . .

A rather unlikely scenario.

A non-linear algebraic equation

http://en.wikipedia.org/wiki/Nonlinear_system

would allow for scaling up or down of the small proportion of "ANE" in Sardinians beyond the rules of linear algebra.

So even if the population of Sardinian men and women had remained relatively stable, a nonlinear effect could then account for the very low proportion of "ANE" in Sardinians, even if R1b carriers were necessarily "ANE" carriers.

Exponentials, logarithms, among many other functions, are examples of non-linear expressions.

Many weakly non-linear systems are modeled using methods of linear approximation.

For example, you can model an exponential (non-linear function )using a Taylor series (linear function). The catch is that it is usually computationally costly to do so.

When we do a PCA analysis on populations, we are in fact using a method of linear approximation to model a weakly non-linear system (human population data sets) with a linear method.

The paper that Chad put up this morning points out that in population genetics, linear PCA analysis sometimes has trouble revealing "nonlinear patterns hidden in a high-dimensional space" (ie they are not dominant effects, but probably should not be ignored if you are trying to discern the difference between two closely related populations.)

Again, I'm not a population geneticist, so I don't know what might lead to non-linear effects in biology. My experience with approximating systems with linear methods is in electronics.

Kurti said...

@Chad and Romulus

Kurds do have significant percentage of I as seen by 2 studies on Iraq and Anatolian Kurds as well individual samples.

But the frequency is probably set too high because of regional sampling.

~10% comes closer to the total frequency.

And Kurds have on average ~4% WHG
So I think this might explain the I?

Marnie said...

@Chad

"it just proves my point that someone can have double digits of a haplogroup and pretty much lack the Mesolithic component that was associated with it."

Yes, I've also notice that.

So hgs are generally only weakly correlated with autosomes.

:)

Chad said...

http://qz.com/336504/a-massive-data-dive-proves-that-languages-and-genes-evolve-together/

https://qzprod.files.wordpress.com/2015/01/alleles-phonemes-map_colorcorrected.jpeg?w=1024

Marnie said...

@Davidski

Thanks for the links. I appreciate it.

Chad said...

Weakly correlated in places of late arrival or only elite dominance. We don't have a Neolithic West Asian to see the turnover, but we do for Europe. There is clearly much more correlation for a turnover with R1a and R1b, as the whole structure of Europe has changed.

Kurti said...

People here are doing so wrong in connecting all ANE just to R*.

How many more facts do people need to acknowledge that a substantial part of ANE is probably connected to J2a and N*. Don't forget the Hungarian Iron Age sample which was autosomally between Ukrainians and North Caucasians and had Haplogroup N*. Don't you think this Haplogroup might have been part of the Indo Europeans? And what about the Hungarian Bronze Age sample which turned out as J2a and had ANE.

pnuadha said...

@marnie You're indirectly accepting public money for your research and in treating the women on your blog differently, you're breaking the law. At least in the US and Canada.

Is this the crap you try to pull at work? Being the only woman in the group does mean that everything that happens to you is because your a woman. Its illogical nonsense which is exactly what most of your statements on this blog are. The constant threats you dish out makes me dislike you personally.

Chad said...

Kurti,
No one is saying that all ANE is related to R, just that R certainly has a deep connection to ANE.

Unknown said...

But Kurti's point stands
Yes ; we cannot directly and explicitly links ane with individual Y Hgs; but we have no R1 in the Hungarian samples whilst we have N and J.

To that David replies there could be no ane in west asia "because the neolithic samples from Europe lack them". Yet; as I've stated; the neolithic europeans could have been from farther South in the levant, as far as nth Africa ; and not northern- highland west Asians

Davidski said...

Mike, there were only two post-Neolithic Y-chromosomes tested in the Hungarian study, and not from Yamnaya remains.

R1a and R1b have already appeared in samples from late Neolithic/Copper Age eastern Germany, belonging to the Corded Ware and Bell Beaker cultures.

Nirjhar007 said...

@Mike
'' Yet; as I've stated; the neolithic europeans could have been from farther South in the levant, as far as nth Africa ; and not northern- highland west Asians''
Of Course.

Chad said...

I've covered this. No sense going over it again.

Unknown said...

Sure ; Dave
I'm sure more will turn up as more LN - EB samples are tested from ECE

Davidski said...

Holy shit, the Near Eastern ancestors of LBK and Koros farmers originated in North Africa and not the Fertile Crescent?

Mike and Nirjhar, no more trolling.

Unknown said...

No; but the Levant was likely diverse itself
And yes; a not insignificant proportion might have come from nth africa - if E-V13 is anything to go by

Chad said...

As far as ANE in West Asia. It probably never crossed the Zagros before 8000BCE, and by the looks of groups like Baden, with Danubian Chalcolithic roots, it might not have happened before 6000BCE. On to the farmers, the oldest sites are in Northern Syria and SE Turkey. Signs point to here as possibly providing the majority of the farmers. If ANE had been in the South Caucasus region, it would've been present at least by those early copper groups. Those people just look like they carried more Basal Eurasian. And how about the "coincidence" of no ANE, with no R1... I'll be damned

Davidski said...

Yes Mike, all of the Near Eastern ancestors of the European Neolithic and Copper Age farmers sequenced to date came from either North Africa or nearby, and in fact probably from the same village. That's why they don't carry any ANE.

Please note sarcasm.

Nirjhar007 said...

@David
''Mike and Nirjhar, no more trolling.''
Its you who is trolling.
@Chad
Is this hard to Understand?
PIE In Zagros-Iran-SC Asia>Maykop->Yamnaya+Antatolia.
SC Asia-West Asian and the around had ANE+R1a+R1b+R2a+Mtdna-U2 from Paleolithic times.
and their are tons of other data too....

pnuadha said...

Weakly correlated in places of late arrival or only elite dominance. We don't have a Neolithic West Asian to see the turnover, but we do for Europe. There is clearly much more correlation for a turnover with R1a and R1b, as the whole structure of Europe has changed.

I think you have the right idea but I just want to point out that based on the information we want we should compare neighboring populations. For example, if we want to know whether or not r1b people brought ANE to western europe you want to look at the distribution of r1b and ANE in western europe and see if they are correlated. The obvious hypothesis is that pre r1b western europeans were similar in their ANE levels but the ANE distribution in western europe became uneven when r1b people, carrying ANE, settled western europe in an uneven fashion.

It is indeed the case that r1b is correlated with ANE in western europe. The apparent influence of R1b in the British Isles is clear; both R1b and ANE concentrate in the Irish/Scottish over the English and French. But the relation of r1b and ANE are less clear when you look at the rest of western Europe. Nevertheless the correlation is still there; r1b and ANE concentrate in N. Spanish, N. French, and N. Italians over the S. Spanish, S. French, and S. Italians. The reason I say the relation of r1b and ANE in Spain, France, and Italy is less clear is because there are multiple ydna's and autosomal dna's that have a similar north south gradient.

Nirjhar007 said...

''Yes Mike, all of the Near Eastern ancestors of the European Neolithic and Copper Age farmers sequenced to date came from either North Africa or nearby, and in fact probably from the same village. That's why they don't carry any ANE.''
Where is Levant and Where is S Caspian? isn't it too close? i think so.
Please note sarcasm.

pnuadha said...

Holy shit, the Near Eastern ancestors of LBK and Koros farmers originated in North Africa and not the Fertile Crescent?

Mike and Nirjhar, no more trolling.


Isn't that what you were doing when you said r1b came to europe via africa?

Nirjhar007 said...

Its still a possibility as Archaeology don't lie and there is the Substratum issue also.....
We need aDNA from Africa to solve the case....

Chad said...

Nirjhar,
I was the one that pointed out months ago on here that there is a connection to South Central Asia, with Maykop and pressure flaking. I gave you that link. As far as what haplogroups are involved, we will have to see. It could end up being female mediated, just as it is in Northern Europe, where we see much more I1 and I2 than G2. If that is an example, we could see R1b and R1a in the Steppes very early. I know there are older forms of R1a and R1b in West Asia. That is not a problem. Who's to say that some didn't go south, into West Asia with the ANE, to have it carried over mostly by females. They could end up not being ancestral to Europeans or only in a minor way.

On to Andronovo, the West Eurasian part of Kazakhs and close groups is just too WHG to not have a good chunk of Yamnaya in it. Expecting to see the average Mesolithic hunters in Kazakhstan at 40-50% WHG is a bit much to swallow.

How about we just wait for papers before declaring any certainties? I'm sure that some on Andronovo will follow at some point in the near future.

Arch Hades said...

I assume this must exclude Britain, with the low levels of R1a and NE European components there's no way IMO that half Britain's ancestry comes from the Steppe.

Nirjhar007 said...

@Chad
''I was the one that pointed out months ago on here that there is a connection to South Central Asia, with Maykop and pressure flaking. I gave you that link. As far as what haplogroups are involved, we will have to see. It could end up being female mediated, just as it is in Northern Europe, where we see much more I1 and I2 than G2. If that is an example, we could see R1b and R1a in the Steppes very early. I know there are older forms of R1a and R1b in West Asia. That is not a problem. Who's to say that some didn't go south, into West Asia with the ANE, to have it carried over mostly by females. They could end up not being ancestral to Europeans or only in a minor way.''
No you weren't i knew about that data i just wanted to know your reference which i also had:).
ANE was there in SC Asia-West Asia from the late Paleolithic chad and I totally Agree With Benedettis Suggestion of PIE in Southern Caspian which also has tons of Materialistic data and more waiting to get published soon and i also find Dienekes's West Asian theory making a strong and similar case.
''On to Andronovo, the West Eurasian part of Kazakhs and close groups is just too WHG to not have a good chunk of Yamnaya in it. Expecting to see the average Mesolithic hunters in Kazakhstan at 40-50% WHG is a bit much to swallow. ''
Just waiting for more aDNA but Andronovo was More Asian than E European at least anthropology wise but we have to know how old is the WHG type ancestry prevailed in S Siberia NC Asian region.
My identification of them is as Pre-Scythians by all means.
''How about we just wait for papers before declaring any certainties? I'm sure that some on Andronovo will follow at some point in the near future.''
Yes but as i say there is always a better a explanation specially concerning the IndoIranian culture and various related data concerning them....

Marnie said...

@Colin Welling

"Is this the crap you try to pull at work? Being the only woman in the group does mean that everything that happens to you is because your a woman. Its illogical nonsense which is exactly what most of your statements on this blog are. The constant threats you dish out makes me dislike you personally."

Wow! You're blaming me for your bullying.

I could care less if you like me or not.

However, I'm not the only one that has noticed that women and others are badly treated on this blog.

The issue is not women versus men or one ethnicity versus another.

The issue is one of a group of well educated members of the public, such as myself, trying to understand the prehistory of Europe, versus a small oligarchy of pseudo academics hiding behind pseudonyms, who are obscuring the truth and following personal agendas, and who attack the person of anyone who dares question their hypothesis.

Davidski said...

Arch,

Around 50% of British and Irish DNA does come from the steppe. I'm sure we'll hear about this after Reich's talk, mainly because it's in Oxford and the people there will want to know. I'll try and post a few leaks from the talk on Tuesday.

Colin,

It seems that cattle and probably people entered Iberia from North Africa during the Chalcolithic, and if so then some R1b might have made the trip too, at least R1b-V88.

But the suggestion that all of the European Neolithic and even Copper Age farmers sequenced to date came from the same location in the southern Levant, and that's why they're ANE-free, is not very smart or useful.

Grey said...

@Chad

"You're not understanding. Megalith people are not running around Britain around 1CE.. They don't matter! Hinxton celts are 2300 years after Beaker starts, 4000 years after megalithic stuff pops up. All that matters is Hinxton and beyond. That is where the change happens."

I understand fine.

If the pre-Saxon pattern was the same as the Saxon one then there'd be an east to west cline with the invaders to the east and the original dna piled up in the west.

So the actual admixture could have happened later and in the west.

Alternatively all the dna from the Atlantic Megalith culture completely disappeared.

pnuadha said...

Wow! You're blaming me for your bullying.

Its not bullying to say your comments are nonsense.

The one trying to bully and unnecessarly intimidate is you. You accuse david of stealing your work and being sexist and capable of being sued. But your outlandish threats are also laughable so...

However, I'm not the only one that has noticed that women and others are badly treated on this blog.


Actually you use your gender to avoid facing criticism. Your always the victim of your circumstances hu?

pnuadha said...

It seems that cattle and probably people entered Iberia from North Africa during the Chalcolithic, and if so then some R1b might have made the trip too, at least R1b-V88.

Sure, why not. But thats not what you said before. You basically said that r1b in the western half of europe is largely the result of r1b entering europe from north africa. I don't see how that is any more feasible than lbk going to europe from north africa.

Unknown said...

Grey,
Ireland is pretty close to the Hinxton folks. They also had megalith people. There is nothing megalith holding out for 2-3k years. There's nothing un-isles like about the Cornish or welsh. Everyone clusters in the same general area. The point being, it is post Celt and Saxon that the shift towards more ENF and less ANE happened. End of story.

Davidski said...

Colin,

In hindsight, it might not be the best explanation, but I wouldn't say it was never feasible.

The Bell Beaker blogger has made a reasonable case for it. Looking at all the data now, I don't think he's right, but in any case there is plenty of evidence of cultural exchanges and perhaps population movements between North Africa and Iberia in prehistoric times.

pnuadha said...

@Mike Thomas

I've told u my argument . But for you its disappointingly agnostic - even nihilistic ; and rather dull. No invading hordes of nomads

Id be very interested in hearing your theories on how ANE made it to western europe, how german CW introduced yamnaya like dna, and the relationship between yamnaya, PIE, the spread of IE in europe, and the spread of IE in Asia.

Also, I agree that the CW migrants to germany carrying yamnaya like dna could have easily be derived from a non-yamnaya source. We should expect a continuum of yamnaya like people in the surrounding area. Its silly to think otherwise.

postneo said...

@chd

"Cernavoda is a mix of Pontic kurgan and varna related folks. There connections back and forth with Anatolia, covered by a couple papers I've seen. One of the later groups, Ezero, does show connections to both the Carpathians and Troy."

This corresponds to Renfrew's out of Anatolia route in the opposite direction. So the paper contradicts Renfrew views in Anatolia whereas for Indo Iranian it is invoked uncritically. How very convenient !

The wheel discussion is fine but is not able to localize to the Steppe. Its horse husbandry alone that is of significance in potentially locating PIE in the steppe. None of the other attributes are of note. Not pastoralism, chariots etc.

Shaikorth said...

Matt,

Not really talking about Baltic Comb-Ware Corded Ware fusion but the groups that later mixed with Corded Ware in Finland, talked about in that Razib post you linked and this article it refers to http://rspb.royalsocietypublishing.org/content/281/1791/20140819.short?rss=1 (re. languages the post-CW cultural mixture inferred from diet is dated to Kiukainen culture, too early to make the groups Uralic speakers even if they were actually Saami-like genetically). LP can be easily selected for, but if the groups were "Saami-like", one might expect lower levels of LP resembling modern North Russians or even farming Volga-Ural peoples. Ancient DNA would be better ofc, LP frequencies can't alone build a solid theory.

Davidski said...

Marnie,

Enough. Do not post any more crazy shit. Act like an adult, or I'll ban you.

Marnie said...

I'm taking the matter Up with Google.

Collin is out of order and you know it, yet you don't do anything about it.

pnuadha said...

@David

In hindsight, it might not be the best explanation, but I wouldn't say it was never feasible.

I see a big shift in your thinking with regards to r1b, but what caused the change. It wasn't that long ago when you made those claims. We already knew quit a bit. We knew(?)...

*That the r1b in Europe is most diverse in the East. There is also a concentration of the precursor to european r1b around the lower danube.

*r1b seems to be missing from neolithic central and western europe.

*r1b likely came to central/west europe during the metal ages, around the same time that IE did.

*Also in the metal ages, WHG increased in central/west europe.

*Finally at some point after the metal ages ANE increased to significant levels forming a fairly smooth gradient East to West.

I don't get what changed your mind. The only bits of info we have gotten recently is that the change in ANE happened during the metal ages, i.e. happened very fast.

postneo said...

@chad

Thanks for the link.

Such phoneme maps are something I have noted long back. Phonemes are super conservative areal features and cut across language families. So far linguists have utilized such data productively.

A curious case: Assamese alveolar stops are found in a sea of retroflexion in south asia. In a symmetrically opposite case scand retroflex stops are found in a sea of alveolar stops in gemanic europe.

These two zones are split by massive zone of dental only stops (middle east, iran, eastern & southern, europe, central asia).

postneo said...

meant to say linguists have not utilized this in productive fashion.

Unknown said...

Colin,
jsut email me on mchaelthomas1980a@ gmail.com
(yes, without the 'i')

and I;ll provide my theoretical account
Its too long to post here.

Davidski said...

By the way, that paper on the R1a from the Altai Kurgans that I mentioned before is freely available here.

http://www.researchgate.net/publication/262806184_Strong_genetic_admixture_in_the_Altai_at_the_Middle_Bronze_Age_revealed_by_uniparental_and_ancestry_informative_markers

Artmar said...

>Marnie, there's also R1a in >Sardinia, including R1a-Z280 >lineages that look like they >came straight from Russia.

Those lineages are mostly Z280 variants close to so called Volga-Carpathian clade (Y2902) and also to some other clades downstream from CTS3402 and CTS1211 (but not identical to them). Also some pre-M458 lineages are inclunded (PF6155xM458 etc.)

Such lineages are today almost not found anywhere else or found in trace amounts, what's really intriguing. Given that there was no documented slavic settlement in Sardinia, those lineages MAY represent a subset od Vandallic lineages related to slavic ones.

>But of course they didn't. They >were introduced from southern >Europe by people who were >usually >typical southern >Europeans in >terms of >genome-wide genetic >structure, >with low levels of ANE.

I guess such lineages were introduced by people considerably similar to modern Central Europeans but they made such low % of society that it just flushed after some generations.

Artmar said...

Sorry for some mistakes, I'm just learning how the comments in this "blogger" system work.

Unknown said...

Excellent study. most of the Y was Z-93. Moreover, shows the population, at least its "west Eurasian" elements derive probably from a largely common pool found from Kazakhstan - sth Siberia , likely to Tarim basin. Question is, how far south this 'common pool' extended. Whatever the case, clearly this pool more or less had its origin in central Eurasia and not the western steppe (David's beloved "Yamnaya")



Dovetails nicely with recent perspectives which see the rise of pastoralism in central and eastern Eurasia as independent to the western steppe.

Davidski said...

Holy shit Mike, did you even bother to read the study?

The Bronze Age Kurgans are very mixed, and clearly of both European and East Asian origin.

The Copper Age individual belongs to Y-DNA Q and mtDNA R, has dark pigmentation, and is probably native to the area, unlike the Europeans and East Asians.

So there was no common "pool" until it formed during the Bronze Age via long distance migrations.

Unknown said...

Perhaps this explains the current distributions
(of course, grossly simplified)

http://imageshack.com/a/img633/8103/AEYLgg.png

Unknown said...

David, I meant to the pool for R1a-Z93 specifically.

And you keep mistaking 'west Eurasian' for European.

Unknown said...

"Long distance".
HHmm - that depends on your definition. But quite possibly the Kazakh- SW Ural - Iran triangle, specifically, from the 'western perspective'. I wouldn't call this area "Europe".

Kurti said...

@Daivd those "Bronze Age Scythian samples" are the once from Altai mountains. They penetrated into now East Eurasian territory what do you think would happen? Exactly what happened with the Tocharians.

Davidski said...

Mike, the Yamnaya horizon was definitely within present-day Europe.

That is what I meant, because that's where all of those ancient R1a groups fanned out from to the east (and west).

Unknown said...

NSS
But you haven't convinced shown that all R1a originated in the Yamnaya region. The (albeit simplistic) map depicted in my link is equally, if not more , parasiminous

Davidski said...

I'm talking about a very specific line of R1a that includes R1a-CTS4385, R1a-Z282 and R1a-Z93, dated only to the Copper Age with full Y-chromosome sequence data.

All other lines of R1a, wherever they might be found or came from, are irrelevant in this context.

Unknown said...

Yes ; those derived from M417
See the simple map I linked ; it's even labelled for u

Davidski said...

That map is best described as an artist's impression, but sadly the artist knew bugger all about the phylogeography of R1a-M417.

For instance, L664 and its upstream mutation, CTS4385, are confined to Northwestern Europe, with no cases ever being recorded in Asia. On the other hand, there are instances of M417* and rather unusual (not typically Asian) Z93* all over Europe, including the steppe.

Nirjhar007 said...

Isn't R1a-417* also exist in Iran?.

Ryukendo K said...

Hi I'm back!

@ Matt
Thanks.

I suspect the same, that WHG was driven up in NE Europe especially, once with the formation of CW, and once again later in post-Iron Age movts, because of the HG continuity through the neol one finds through paleoanth and also through the other pieces of evidence brought here by others, which indicates to me that Yamnaya immigrants would have had to interact with these pops when they arrived.

@ Shaikorth @ Davidski

Thank you for your comments.

Agree that ENA was a possibility, just that there are counterargs. It would be obvious in D-stats with Dai in between Karelian and Lochour/Mal'ta, which if the paper does not try doing we can here.

In any case, the authors said Yam 'can be fitted as' Karelian+Armenian, a result probably from a an algorithm, not necessarily that it was literally, physically Karelians + something like Armenian, which is something literally impossible to prove. Like how f-stats and THREEPOP chooses the 'best' mixtures for present-day pops. I bet that EHG is a ghost pop that included but not identical with Karelians, since thats usu. the way they go about doing things.

Shaikorth, I would really like to hear what you think about the technical issues.

@ Kristiina

Thanks for your comments.

Moksha are basically identical with finns in ANE and ENA, but different in WHG and ENF, which makes it impossible to model finns--on the basis of ENA--as Moksha + local, as that would require them to be 100% Moksha which would throw off WHG and ENF. Also, Moksha have more ENF than finns, which makes me not trust them as proxies for Uralics for obvious reasons.

I believe Saaami are also sufficiently low in ENA as to also require >80% contribution to account for ENA in finns, which would throw things off too. This seems to be a common theme in these kinds of estimates, e.g. attempts to fit NW Euros as poles + local force contributions of >80% from poles.

This makes the likely scenario that another pop contributed to both Finns and Moksha, in a scenario remniscient of CW in Europe, and the closest I could find was Maris, who, incidentally, live in the Volga-Ural region, close to where Uralic in Eur is believed to have come. They are 25% ENA and 22% ANE.

Ryukendo K said...

@ Marnie
It seems like more people than just you feel that WHG, ANE and ENF are somehow not rooted in solid pop gen understanding. There's a lot to explain, but let me just give you the most obvious pts.

When the scientists uncovered the genomes, it became immediately clear that they were outside of modern genetic variation in Eur, but also--more importantly--that the largest fraction of variation in modern genetics in W.Eurasia fitted in the dimensions where these aDNA are most distinct from each other.

Prior to the ancient genomes, PCAs would always produce the familiar genetic picture with the largest two dimensions as Kalash vs. Sardinian, and Estonian/Lithuanian vs. Bedouin, and it was not clear at all why this was the case. Similarly ADMIXTURE of west Eurasia always give Bedouin vs. Estonian/Chuvash components for West Eurasia at K=2, and this was utterly mysterious.

The aDNA was tested against modern pops via D-stats, which are highly robust, to produce the Basal((ANE)(WHG)) model we are familiar with today. More importantly, this cast light on literally everything. All of a sudden it became clear that high Crown Eurasian differentiated Estonian/Chuvash from Bedouin, and that the largest dimension in PCA was dominated by Basal vs (WHG-ANE), and the second largest by WHG vs. ANE.
When the aDNA was fitted into PCAs at precisely the right locations, WHG in the northwest corner, ANE at the Northeast, Basal at the south-center, was that much more confirmation, and proved that where these genomes differed from each other was also where the largest fraction of structure among modern W.Eurasian pops would also differ.

Of course there are higher dimensions where later, less impt differences are found, e.g. between Armenian and Sardinian, between Irish and Balts, etc. but that they are higher dimensions mean they account for much less variation by default.

So when David uncovers what he does, he can be pretty sure that the things he talks about are real. Statistical phenomena have no less physical reality that binary phenomena.

Last of all, let me remind you that the world is a pretty libertarian place, its the political ideology that is closest to what happens in practice in every place, everytime, and thus the personal is not the political pretty much anywhere outside the most academic circles, and attempting to make it so helps neither you nor your cause.

Shaikorth said...

RK,
formal testing will tell if EHG has extra ENA affinities compared to a WHG+ANE mix. I'm pretty confident Reich and co going to discuss it more in the paper, while the presentation and abstracts so far were more about Yamnaya/CW and their composition. One question are if the Samara HG's are interchangeable with Karelians as EHG proxies, and if they are why they named Karelians as proxy since the Samara HG's are actually from the steppe (and more likely to mix with whatever was coming from the south to form Yamnaya) and not from distant forest zone.

Re. Uralic issue, there needs to be ancient DNA from an Uralic urheimat or something to settle that, and I don't think there are good surviving proxies. Maris for instance have some very recent Turkic admixture from historical times as per Yunusbayev paper and possibly something else too. The more eastern Baltic Finns (which aren't included in K8 sheet) just become harder to model as something like "Estonian + Mari", while Lithuanian + Chuvash/Tatar works for Kargopol/Erzya/Moksha. Very visible on for example this PCA where dimensions are defined by David's ANE component formed from non-ENA parts of Karitiana.

https://drive.google.com/file/d/0B9o3EYTdM8lQUE9zUVpjRnUtMFU/edit?pli=1

Ryukendo K said...

@ Various

On the topic of the specific dynamics of Y-DNA, rmb that male coercion in wartime and female choice in peacetime and higher effective fertility for elites almost always result in rather disproportionate success for high-status males over time.

The Y-haps of mexicans are, like, >85% European? Despite being 50% European autosomally. Its not very pc to say this, but one glance at mexican adverts or telenovellas kinda highlights the reason...

Figures off memory, approx only.

Economic historians have more or less shown that status is highly heritable in an extremely wide range of both ancient and modern societies for father-son comparisons, thus it makes sense that a slight success for high status males per generation in either 1) number of wives or 2) number of survivng children or 3) the status of his sons, would result in a very large disproportion over time.

Prob 2 + 3 in settled agri societies and 1 + 3 in less peaceful times. Which actually seems to be to be borne out by comparisons with the amount of discrepancy w.r.t. autosomes. Agri elites result in both disproportionate autosomal and Y-DNA contribution, but Y-DNA still more than autosomal, e.g. Madagascar, Mexico, Brazil, South China, England vs Ireland/Scotland, thus 2+3. While very early elites in the metal ages where polities are still being formed seem to result in disproportionate Y-DNA and much less dispro autosomal, thus 1+3.

Artmar said...

@Nirjhar
I haven't heard of any genuine M417*, especially in Iran.


@Davidski, I hope that my info on Sardinian R1a haven't passed unnoticed.

Ryukendo K said...

@ Shaikorth @ Davidski

Thanks for that figure. I agree with your points.

Something that that ANE figure highlights though, is that the weird position of finnish w.r.t. ANE can in no way be acccounted for by admixture from less diluted uralics further east. A vector from any Europeans to Finnish does not point at Uralics in that PCA; rather finnish form their own extreme which they share with balts.

Since both Yamnaya and CW will score south of them, something pulled finnish, balts, and probably some other E.Euros north. This just seems plainly obvious to me.

Shaikorth, what do you have to say about my comments regarding ADMIXTURE?

Ryukendo K said...

@ Kristiina

I think your comments regarding microblade technology are incredibly interesting, as the other person to make this connection is... Edward Vadja. I suspect you know this already, but for the benefit of the rest he's the guy who proved Dene-Yeniseian, who is the only guy to have done any very groundbreaking work with the actual comparative method instead of exotic 'long-range comparisons' in linguistics, and the connection was between Na Dene and the language of the Kets. Its quite robust and is supported by most linguists now--the newest language family!

And breaking the time barrier by an unprecedented degree too. Its prob >15000 ya.

He speciffically made the connec. between Hap Q, ANE and microblades. If I were to make a guess the spread seems to correlate with the emergence of SHG and EHG as mixed pops in Europe.

Nice to have some linguists boarding the anthro train so enthusiastically.

Signing off for the night.

Shaikorth said...

"To summarise, in the same way that ADMIXTURE can overlook/be insensitive to time issues and IBD/IBS and use collinearr drift in the direction of Mal'ta to produce estimates of ancestry from Mal'ta in the presence of other clusters with collinear drift for population groups at the same 'taxonomic level', aka WHG and ENF, ADMIXTURE primarily uses collinear drift to target even later ancient genomes and divvy up their ancestry in the presence of unsupervised clusters, no?"

I'm going to say this is possible. Hopefully someone is going to run a paper on ADMIXTURE and issues of ancient formally unmixed and modern genomes unsupervisedly together with some extensive testing. For now we see practical effects and, based on that, make some decent predictions. Ancient will be split into modern components, differently depending on both type and number of modern samples in the run. Intentional sampling can affect it, but it's a balancing act since reduction of samples results in missing variation. If there are enough high quality ancient genomes in the run, they'll start forming unmixed ancient clusters and modern genomes start to appear mixed so the problem may just solve itself with time as researchers get more samples.

Marnie said...

@ryuk

"Last of all, let me remind you that the world is a pretty libertarian place, its the political ideology that is closest to what happens in practice in every place, everytimeI Last of all, let me remind you that the world is a pretty libertarian place, its the political ideology that is closest to what happens in practice in every place, everytime, and thus the personal is not the political pretty much anywhere outside the most academic circles, and attempting to make it so helps neither you nor your cause."

I don't have a "cause".

Anyway, my time is probably better spent working on filing the formal documents to find out who you guys are, and formally complain to your institutions about your anonymous use of blogs and message boards to strong arm members of the public.

Ryukendo K said...

@ Marnie

Gawd Marnie, I was trying to be polite to you. I guess I just came across as snarky and oppressive?

Jeezus!

Marnie said...

@ryu

No, you didn't come across as snarky at all.

I appreciated your technical explanation.

Overall, though, it's the ANE, WHG, farmers model is OK, as I've stated before, with a few caveats.

Again, I don't have a cause, have no particular convictions at all, and if someone could prove to me beyond the limits of uncertainty some of these ideas you guys have, I'd happily accept them.

I really don't care where Europeans (or anyone else) came from.

But you haven't proven that, and there are several significant flaws in these models, which you haven't reasonably accounted for.

Anyway, it appears that anyone who tries to point these things out is smeared. That's the real problem.

So, again, I won't be blogging here anymore, and I've "quarantined" the research from all of you, not because your models are likely to be incorrect, but because some of you (Colin in particular) are bullying anonymously. That kind of behavior seems to be acceptable to your group.

Nirjhar007 said...

@Artur
''I haven't heard of any genuine M417*, especially in Iran.''
5 Folks of Iran were found to be belonging to R1a1-417 SRY10831.2* According to Underhill Paper of 2014.
Can they mean something else?.

Simon_W said...

On a second thought, BR1 may yet be similar to Yamnaya. I wrote that she had hardly any West Asian admixture, but that holds true only for the Eurogenes K15 analysis. BR1 isn't on GEDmatch, so I can't analyze her with other calculators, but the very similar BR2 is. And BR2 is 10.1% West Asian according to Dodecad 7b, similar to English people. This West Asian component was also present in NE1, but not in the samples inbetween, suggesting a renewed arrival of it. The Eurogenes K15 West Asian component appears in IR1 for the first time. And IR1 is 24.2% West Asian in Dodecad K7b, that's a lot, similar to the Greek level. Apparently it takes really serious West Asian admixture for it to become appreciable in Eurogenes K15. Now we don't know how the Yamnaya individuals will score, but at least BR1 wasn't entirely without West Asian admixture, it's just included in her other Eurogenes K15 components. E.g. the Baltic component, based on present-day people from the Baltic, may easily include some West Asian admixture. Therefore the EHG ancestry in BR1 and BR2 may be from the local Yamnaya, and in that case we can assume an IE language for them.

This may have been Italo-Celtic, in line with Anthony's theory. In that case Celtic spread to northwestern Europe rather late and with a rather thin elite. The greatest deal of the ANE on the British Isles is probably older, stemming from central European Bell Beaker people. I'm assuming this because BR1 just hadn't enough ANE.

Or, other hypothesis, pre-proto-Italo-Celtic somehow spread from West Asia to Southwestern Europe with R1b and spread with the Bell Beakers to central Europe, where they assimilated ANE-rich EHG descended Corded people, giving thus rise to the typical Celts with their northern DNA.

I'd really like to know which hypothesis is correct...

Shaikorth said...

BR2's West Asian seems to be "hidden" in East Med using Eurogenes K13 and K15. Eurogenes Hunter-Gatherer test picks it up as Anatolian Farmer, which peaks in Caucasus. Anatolian Farmer (peaks in Caucasus) and MDLP World-22 shows it as West Asian, which is a bit higher in BR2 than in modern Swedes.

It doesn't necessarily mean BR1 and BR2 are Yamnaya-like, but they likely do have some mixture from a population like it.

Nirjhar007 said...

@Shai,Simon
Which most ancient European aDNA shows significant presence of West Asian from where and which date?.

Grey said...

@Chad

"End of story."

Okay, all the dna from the Atlantic Megalith culture completely disappeared until your "continentals" (mix of celtic and pre-celtic) arrived post-Saxon from Europe...

until there's wider coverage and it's shown that this continental mix of celtic and pre-celtic was already in Britain pre-Saxon and concentrated in the west exactly where you'd expect to be if the Celtic invasions followed the same east to west pattern as the Saxon one.

Grey said...

@Chad

"it is post Celt and Saxon that the shift towards more ENF and less ANE happened."

I don't dispute that bit btw. I'd put it during the industrial revolution when the east and west halves mixed in the cities.

Shaikorth said...

Nirjhar, it's the Iron Age samples if we define "significant" as in modern European context. BR1 and BR2 scores are only significant compared to a few populations like Basques, Finns, Sardinians.

Kostenki14 has a funny Eurogenes K15 score, at least a little of every component except West Asian and East Med. As I expected from that, it showed 0% Anatolian Farmer in the H-G test too.

Nirjhar007 said...

Who has the highest West Asian?.

Shaikorth said...

IR1 easily.

Nirjhar007 said...

Thanks!:).

Grey said...

@Simon_W

"A major problem for the theory of a steppe origin of PIE is: The steppe foragers have adopted the West Asian way of life as herders. Whose language was more likely to prevail? Of course the language of the dominant cultural model...I can't see how the hunter-gatherer language should have prevailed under such circumstances."

At the edge of the farming zone beyond which farmers physically couldn't spread (cos climate or whatever) you'd get a border between farmers and HGs.

If a hybrid HG&Herding culture was viable in the HG zone then the language that ended up being dominant in that zone would depend on which group adapted to that hybrid model first.

Take Arabia as an example with HGs adapted to the desert and farmers rolling up to the border in their armored ox cart divisions.

Is it more likely the farmers adapt to surviving in the desert faster than the desert HGs pick up herding (possibly via rustled animals and captives)?

It could be either but I think HG first is more likely in that specific kind of situation.

This doesn't prove anything about PIE. It just shows how it might happen.

.

"Especially if the West Asians made up 50% of the population."

This would weigh against the previous argument unless it wasn't 50% at the beginning and only became so later.

.

One possible explanation for where a desert-like condition in the north may have come from might be the range of the Boreal forest at the time as Boreal forests have thin and acidic soils so farmers cutting down trees and planting alkaline soil adapted crops might not work whereas burning a chunk of forest and feeding cattle and pigs on the regrowth might work fine.



Grey said...

Current Boreal range

http://www.marietta.edu/~biol/biomes/images/taiga/taiga_500.jpg

but where was it 4-6 thousand years ago when it was colder?

nb you also get boreal forest in high altitude (colder) regions outside the boreal zone so it's a cold climate biome with plants and critters (and possibly people) adapted to that biome.

Matt said...

Krefter: Hunter gatherers were living up north ~4,000-5,000YBP, not in the Balkans.

Hunter gatherers were living in Germany around 5000YBP - http://phys.org/news/2013-10-european-hunter-gatherers-immigrant-farmers-side-by-side.html

Probably they were living in Spain and the Balkans still as well.

However, their relative size compared to the farmers would be lesser in Spain and the Balkans, simply because there were probably more farmers in Spain and the Balkans, where Near Eastern agriculture worked better than Germany (and in Germany Near Eastern agriculture worked better than in the Baltic). So any contribution to the larger Spanish or Balkan population would be of lesser effect (like sending 100 European to an isolated village of 500 Africans vs 100 Europeans to a village of 1000 Africans - less impact on the population).

Krefter: It seems you think Yamna-type ancestry largely spread independently in each region of Europe, and I dis agree.

It seems like what Reich's new paper may show is that early Neolithic farmers spread out quickly without intermixing. For them it was not a "rolling snowball" of picking up HG ancestry along their path as you described in your post. Then slowly began to intermix with local populations later in history, while continuing to breed somewhat with one another.

I don't see that this is necessarily improbable for Yamnaya / Corded Ware.

capra internetensis said...

Tocharian is attested very late, right on the main route across Eurasia. It could have washed up there off the steppe from pretty much anywhere at any time. It doesn't need to have been in the Tarim Basin for very long.

@postneo

I'm pretty sure the real point of that paper is to say screw you to Gray and Atkinson.

It isn't invoking *Renfrew's* theory of the Indo-Iranian spread at all. Renfrew offered 2 separate hypotheses for the origin of Indo-Iranian - a spread eastward with the West Asian Neolithic, or a spread from the steppe, ultimately originating with Cucuteni-Tripolye farmers who adopted pastoralism and the pit-grave culture. He included the latter scenario, in opposition to his primary theory, only because of the difficulties with the first hypothesis - if Indo-Iranian had spread east across Iran, then not only do we have to explain how it was then replaced by Sumerian, Semitic, Elamite, and umpteen other languages in between, but also why it isn't highly divergent from all the other Indo-European branches, having been separated for the longest.

But that steppe spread hypothesis goes back much further - already Gimbutas in 1963 had placed Indo-Iranian origin in the Andronovo culture. The recent and thorough exposition is Kuzmina's book "The Origin of the Indo-Iranians".

So the hypothesis is not originally Renfrew's, and his version differs sharply from Anthony's.

That is not to say that I think this theory is necessarily correct, of course. A spread from south of the Caucasus or Caspian has certain advantages.

Davidski said...

Artur,

I remember reading something about Slavic mercenaries fighting for the Romans in Sardinia. It's all laid out in a book on Roman history that can be previewed on Google, but I forgot the title and author.

Unknown said...

" Uralic issue, there needs to be ancient DNA from an Uralic urheimat or something to settle that, and I don't think there are good surviving proxies. Maris for instance"

That would be great . But depends where u think u need to go looking for the Uralic homeland .

@ Capri

"
Tocharian is attested very late, right on the main route across Eurasia. It could have washed up there off the steppe from pretty much anywhere at any time. It doesn't need to have been in the Tarim Basin for very long time "

Glad to see the occasional intelligent blogger here

Unknown said...

Davidski
Sklavene mercenariesbrought throughout italy for the byzantinesndurong the gothic wars. They also fought in Asia minor and as far as Syria

Unknown said...

Davidski
Re: R1a
You fail to provide a convincing account for the origin of R1a. Stop pontificating and show convincing data ; if so, then I'll gladly accept and praise where it's due

Davidski said...

Mike, it's not important in this context where R1a originated. What really matters is where and when the three main lineages of R1a that exist today - R1a-Z93, R1a-Z282 and R1a-CTS4385 - expanded from.

Based on complete Y-chromosome sequences it's reasonable to say that all three lineages expanded very rapidly at about the same time, which was probably after the Neolithic. Indeed, this gels with ancient DNA evidence, which picks up no R1a until after the final Neolithic, and then it picks up a lot of it in north Eurasia and also the Tarim Basin, and always associated to some degree with European-specific mtDNA sequences and also mixed eye/hair pigmentation.

So obviously this information fits very well with the PIE steppe hypothesis, especially if we also consider that these markers are all closely associated today with Indo-European language groups like Baltic, Slavic, North Germanic and Indo-Iranian.

Turks and some Uralics also carry a lot of R1a-Z93 and R1a-Z282, but this doesn't mess anything up. In fact, it fits right in, because the early Turks are thought to have absorbed the earlier Indo-European peoples of the steppe as they moved west from the Altai, and Uralics were always close neighbors of the Indo-Europeans, since proto-Uralic times.

Unknown said...


Sure; a recent expansion thru Western Eurasia cannot be doubted ; also I sympathise with your arguments based on distribution; and its correlation, at least at face value, with *some* iE groups .

I guess, at least for me, the real clincher would also be unravelling the history of R1b and excluding/ proving a pre-Neolithic ANE presence in places apart from the western steppe.

If those conditions are met; then parasimony would very much be in the western Steppe court. And as I said, I agree with the chronology of the kurgan hypothesis ; at least ; for its approximate beginning .

As stated earlier; I'd find it a historical miraculous feat if it really did spread at the rapid rate & geographical extent at which the traditional axiom argues

Unknown said...

I should temper that by saying that applies for the current, i.e. extant, branches of R1a. It is possible that the current branches merely overlay older now extinct branches of R1, and/ or other Hgs, which are associated with the initial peopleing of northern and western Eurasia .

Davidski said...

Well modern genomics agree with the traditional axiom. It was a very rapid spread.

Nirjhar007 said...

@David
'', it's not important in this context where R1a originated. What really matters is where and when the three main lineages of R1a that exist today - R1a-Z93, R1a-Z282 and R1a-CTS4385 - expanded from.''
Pathetic.
''Turks and some Uralics also carry a lot of R1a-Z93 and R1a-Z282, but this doesn't mess anything up. In fact, it fits right in, because the early Turks are thought to have absorbed the earlier Indo-European peoples of the steppe as they moved west from the Altai, and Uralics were always close neighbors of the Indo-Europeans, since proto-Uralic times.''
More Pathetic.
''Well modern genomics agree with the traditional axiom. It was a very rapid spread.''
The Most Pathetic.
Clear case of One Eyed Interpretation.

Davidski said...

Postneo,

Anthony and Ringe didn't copy Renfrew.

What happened was that Renfrew's first model was so obviously wrong that he had to modify it significantly, and ended up copying a part of the classic steppe PIE model.

You should know this sort of stuff.

Ryukendo K said...

@ Simon W @ Shaikorth

Agree. It seems to me quite apparent by now that Yamnaya and CW flows correlate to a period of increased influence from ENF from the Near East directly.

Its been immensely frustrating trying to figure out a formal way to prove this though. ADMXTURE and other model-based algorithms are too sensitive to what pops we include in the dataset and the vagaries of K and such, so can give us strong indications of the sources of the components and not just the ratios but prob won't be accepted as proof by some ppl, not entirely without reason.

If it were possible to find out where the ANE, ENF and WHG in Yamnaya derived it would be illuminating indeed. I suspect an origin for both the ANE and WHG portion from far to the east, and ENF directly from the near east, but it seems impossible to prove this with the tools we have.

The closest I can come up with is a f3-shared drift stat with Yamnaya zombies made by adding Mal'ta to Basque and to Lezgin until the ratios match, but this method is far too messy; also, when David tried this with his PIE frankenstein the best source of ANE for Euros in f-stats was still Karitiana. Other methods all seem like they would be hijacked by the differing levels of the three components, not the sources, in the samples we test. So still pulling blanks on this one.

«Oldest ‹Older   401 – 600 of 646   Newer› Newest»