search this blog

Friday, January 23, 2015

Yamnaya genomes are a 50/50 mix of eastern Euro foragers and something else ANE-rich


I'm posting a new entry about the upcoming Corded Ware/Yamnaya paper because the last entry (see here) now has over 400 comments which aren't easy to load for many people.

One of the authors of this eagerly awaited paper, Nick Patterson of the Broad Institute, briefly joined our discussion. Nick's contribution is much appreciated. He wasn't able to reveal a great deal, because the manuscript is in submission, but he did make a couple of interesting points:

- the paper will feature Y-haplogroup results from the Yamnaya culture, represented by nine samples in all, including seven males

- the population with Near Eastern ancestry that mixed with the Eastern Hunter-Gatherers (EHG) on the Russian steppe to form the Yamnaya pastoralists by 5,000 YBP was also "rich" in ANE

- ancient DNA from the Caucasus, Iran and India is probably necessary to work out how the Indo-Europeans got to India, but the paper won't feature such data

It's nice to hear that Y-haplogroups aren't being ignored. My opinion is that they're at least as important as genome-wide data when tracking the movements across vast space and time of highly patriarchal and patrilineal groups like the ancient Indo-Europeans.

Indeed, we already know that the Slavic, Baltic and Norse-specific R1a1a1b1, defined by the Z282 mutation, is the sister clade of the Indo-Iranian-specific R1a1a1b2, defined by Z93. Thus, if the Yamnaya males were found to belong to these or upstream markers, this would suggest that they were the paternal ancestors of many Balts, Scandinavians, Slavs and Indo-Iranians, and correlate very nicely with the linguistic and archeological "steppe hypothesis" of Indo-European origins.

In fact, even if analyses based on high density genome-wide data suggest that Indians don't harbor any genome-wide European ancestry, we'd still have to accept the likelihood of gene flow - albeit perhaps very indirect gene flow - from the European steppe to India because many Indians belong to R1a1a1b2.

The second point made by Nick is perhaps surprising, but at least for me not totally unexpected. That's because we've already known for a while that the Yamnaya genomes can be successfully modeled as half Karelian EHG and half present-day Armenian (see here), and according to my own estimates Armenians carry an average of 15.5% ANE.

The fact that these Armenian-like, ANE-rich newcomers dampened the genome-wide affinity to ANE-proxy MA-1 on the Russian steppe might look like a contradiction, but not if we remember that the higher the Near Eastern ancestry the lower the genome-wide affinity to MA-1, and also consider that the steppe foragers probably carried a lot more ANE than the newcomers.

Actually, as far as I know, all of the Yamnaya samples in this study come from the Samara Valley, which is some distance north of the Caspian Sea near the southern Urals. So it makes senses that the pseudo Armenians who turned up there more than 5,000 years ago were not like the Neolithic farmers of Western and Central Europe, who lacked ANE.

I'd say that this as yet unidentified group (wild guess: immediate ancestors of the Repin culture people?) was the result of an admixture event, or perhaps a series of admixture events, with ANE-rich foragers somewhere on the steppe south of the Samara. If so, I won't be surprised if it turns out that R1a only appeared in the Samara Valley after their arrival.

In any case, it looks like even after this paper comes out, we'll still need a lot more ancient DNA from across Eurasia to help map out the early Indo-European dispersals with any confidence.

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

499 comments:

«Oldest   ‹Older   201 – 400 of 499   Newer›   Newest»
Nirjhar007 said...

@Maju
''"Drift", "luck"... I'm bored of reading those spurious arguments. They are just trenches of the mind, just like "faith".''
But its surely a possibility.
''Well, something: lineages, autosomal DNA fractions, you know... They have nothing that suggests they are from South India, nothing that makes them different from Balochs particularly or their other neighbors more in general.

So the immigrant conjecture does not fit: they are living proof of Dravidian being spoken in Pakistan long ago.''
What about their South Indian Component how that scores compared to other surrounding populations???

Unknown said...

"How is that not relevant? You can't assume Poles and Germans have similar ANE, when they obviously are very different genetically. Germany being very diverse as a country is also relevant, because they don't all have the same ANE."

on the other hand, the homogeneity of Poles, despite being geographically as large (more or less) as Germans, should alert us against using them as a proxy for some undilated CWC-derived ancestors.

Krefter said...

""Drift", "luck"... I'm bored of reading those spurious arguments. They are just trenches of the mind, just like "faith"."

This may be off subject, but I'm responding to statements you've said about mtDNA.

Drift is a very real thing. Lithuanians have close to 20% typical Mesolithic Euro mtDNA, but have 75% Mesolithic Euro blood. Drift, gender-biased admixture, natural selection, and what not are the reason.

mtDNA H is not the marker of a group of people defined in autosomal genetics(like U5), as far as we know. If anything it can be associated with ENF and basal Eurasian.

Marking a migration of "H", a very very old maternal lineage, is simplistic. There was no pop in Europe in 5,000BC who had something like 80% H.

If a pop from a single region caused H across Europe to rise from 25-30% to 40%, where are their autosomal markers and non-H lineages? Ancient DNA is suggesting that after the Neolithic gene flow in Europe mostly went from east-west not west-east.

I'm not saying it is impossible H has little to do with drift, but I doubt it.

Old genetics association of H with Europe was overplayed. It's bad research, like when the Minoans were called "European" based on mtDNA. IMO mtDNA is prone to drift, and what matters most is "most recent common ancestor", not what clade frequencies.

Unknown said...

I agree krefter. We cannot overlook drift, selection on uniparental markers , esp given their small Effective size compared to autosomes, and function in reproductive success and metabolism, etc, respectively.

Krefter said...

Mike and others,

Is it possible that drift autosomes, can make autosomal genetics unreliable? It's mythical that we can get a very good idea exactly what pre-historic modern people descend from and at what percentages.

I'm wondering this because Euros have overwhelmingly near eastern mtDNA, yet many are closer autosomally to Mesolithic Euros than near easterns. But then again ancient mtDNA and autosomal DNA from Europe, are giving answers as to why mtDNA turned out to be mostly near eastern, while overall ancestry stayed very Mesolithic-like.

Krefter said...

Also, K-14 looks ancestral to WHG, and maybe ANE.

What are the chances for that? It can't be random he shows the same relation to modern pops as WHG-ANE, after 30,000 years or whatever of mutations. Models created for modern ancestry based on 8,000YBP genomes are consistent with 38,000YBP genomes, which is amazing.

Nirjhar007 said...

@Krefter,Mike,Maju,David-
Check this one Armenian Origins-
http://dienekes.blogspot.in/2015/01/dna-origins-of-peoples-armenians.html
Fascinating stuff.

Maju said...

@Nirjhar, Krefter & Mike: arguing that water becomes fire by mere chance at your convenience is not a possibility: it's magic, it's a miracle, it's a fairy tale! Blaming "drift" for dramatic major changes in population genetics is like saying "hocus pocus" or "God/a wizard did it".

It's not a serious scientific argument.

Of course drift does exist but do you guys really understand how it works? First of all, the larger the population and also in expansion contexts, drift tends to become negligible. Drift basically tends to fixate the dominant aspects of the genetic pool, so you can beging with 60% A, 40% other and end with 100% A. Only when the populations are very small and stable drift can become (normally) a matter of real chance, for example when Ne=2 or in general Ne<10, when Ne=100 (all are arbitrary but realistic figures) the random aspect of drift becomes almost irrelevant except for the tendency to fixate what is already dominant and even that tendency slows down quickly as Ne becomes larger.

Just do the maths. Or if you can't get a math oriented friend or colleague and ask them to help you with this purely mathematical problem.

So drift happens in New Guinea, Lapland or among small hunter-gatherer populations but does not really happen in larger effective populations. Major changes in these populations demand other explanations, typically external colonization, often paired with genocide.

Not a pretty answer, I know, but quite realistic.

"Euros have overwhelmingly near eastern mtDNA"...

Another nonsense. MtDNA H is not "Near Eastern" at least not anymore than European. This has been demonstrated empirically but people like you and some others continue entrenched in such absurd ideas.

Europeans do have a sizable chunk of West Asian mtDNA, namely: J, T, X, W and almost certainly also K (which is part of U) and N1. And they do have a sizable chunk of West Asian + African Y-DNA: E1b, G2a, J2b, etc. Very roughly in the same amounts as they have West Asian autosomal DNA. The figures add up if you look at the problem with the right lenses but not if you do with extremely distorted ones.

The problem is not in "drift" but in your a prioris, and in the mental trenches you dig to defend those a prioris.

Nirjhar007 said...

Maju, since Linguistics agree for a 1000 ad origin drift is suggested by applying common sense anyway what is the ASI structures score in Brahui people?

Krefter said...

I sent an email to Patterson asking about Late hunter gatherers from eastern Europe mixing with Yamna along with EEFs to create CWC, and this is what he said.

"I think you will have to wait until we release the paper, which I hope will be fairly soon."

It seems he's saying he and his peers are in control whether the paper is released, not editors. Could this mean they might put the paper on bioRxiv?

If my question doesn't make sense, it's because I have no idea how the academic world works.

Krefter said...

"Another nonsense. MtDNA H is not "Near Eastern" at least not anymore than European. This has been demonstrated empirically but people like you and some others continue entrenched in such absurd ideas. "

H+U5+U4+U2=~60% for typical Europeans. Even if you think all H is of Mesolithic origin, that's unproportional for many Europeans.

Lithuanains and south Italians have similar overall haplogroup frequencies but very different autosomal DNA. So, you're philosophy of haplogroup frequencies being better for testing relatedness than "Most recent common ancestor", doesn't work for those people.

If H did exist in Mesolithic Europe without sometype of near eastern-intrusion(which I think it very well may have), it was very very rare.

Maybe it was very popular in Iberia, I don't know, but that would be strange considering their close relation to Mesolithic people who lived north of them. But Mesolithic Iberia can't explain similar H-clades throughout Europe today and in Neolithic central Europeans.

If H existed in Mesolithic southeast Europe, those people probably had ENF-like ancestry.

Unknown said...

But maju; I've already explained that there were and have been numerous population crashes , albeit regionally circumscribed , since the Neolithic . It would be foolish to envisage a steady state growth since 7000Bc

Krefter said...

"The problem is not in "drift" but in your a prioris, and in the mental trenches you dig to defend those a prioris."

You're the one who's been year after year of Mesolithic Euros coming out vast majority U, insisting H was popular in Mesolithic Europe. After dozens of Y DNA samples from Neolithic west Europe and no R1b, you insist R1b was there.

When it was discovered Mesolithic Euros are closest to northern Euros and Neolithic Euros are closest to south Euros several years ago, you insisted there were Mediterranean-type pops somewhere in Europe during the Mesolithic.

So, you have an obvious Basque-centric streak. I'm not the only one who's noticed it.

Krefter said...

Maju, you have to deal with what the evidence says: If there was H in Mesolithic Europe, it has few descendants in Europe today.

postneo said...

Can someone please explain if ANE itself is a south caspian component meaning its a bad label.

We cant expect modern day armenians, iranians and kurds to have the same ANE levels as before since they have been increasingly influenced by levantine populations.. genocide even. If you go east of Iran to more sheltered areas what happens to ANE % ?

Krefter said...

Maju, I'm fn tired of your attitude.

Keep your intolerance and resentment for religious people to yourself(and hopefully you'll change in that way), and save discussions about "faith" and "God" for times when they're relevant.

Having a productive discussion and or argument with someone in a forum or blog does not include insulting them in every other post.

Continuously Dis respecting and insulting someone because they dis agree with you does not classify as friendly(we are not your enemies) social behavior, even among close friends and family members. I'm not saying you should be friends with people online, but you should not be their enemy.

Even if you're right, as long as you dis respect the person you're trying to convince, they'll only become bitter enemies with you.

Davidski said...

ANE isn't a south Caspian component. It really couldn't have arrived near the Fertile Crescent before the early Neolithic.

It seems to peak in refuge areas like the northeast Caucasus, Hindu Kush and also central Siberia among the Kets and Selkups.

Here's a spatial map of ANE...

https://drive.google.com/file/d/0B9o3EYTdM8lQOS1YODhvOWpWNzg/view?usp=sharing

Unknown said...

David you're probably right
But also possibly wrong
Neolithic Europeans might have been Levantine - Southern near easterners rather than northern- highland ANE rich ones
Certainly, the evidence for E-V13 points to a Sth levant - North African source

Unknown said...

In fact , I posit multie sources for ANE shift in post-Neo Europe
1) some Palaeolithic resurgence - to use a journalistic phrase
2) northeastern R1a -associated; more likely within eastern - Northern Europe itself rather than solely from the steppe
3) highland west Asian ; R1b associated . Possibly the dominant source

My hat is up for eating

Kristiina said...

In my opinion, you tend to attach far too much importance on y lines as factors of genetic divergence and convergence. Genetics probably gradually change, for example, when, for climatic/geographical reasons, new passages open up or new areas become available to settlement or when new cultural focus areas arise and there is a continuous genetic exchange between new groups. The new ways of transport were surely increasing this change. My feeling is that the more ancient y DNA is coming out, the more I think that y lines may come and go very rapidly while people do not change so much due to these changes in y lines. East Eurasian and West Eurasian seem to have existed at a very early date when it may have been yDNA C in Europe as well as in East Asia. In Finland and Baltic area, most y lines seem to be young and vital new y lines that brought the new inventions and new ways of living to the north.

Finns and Balts share the same genetics and haplogroups as their histories are intertwined from the repopulation of northern Europe onwards. They share the following cultures: Kunda culture(8000–5000 BC), Comb Ceramic culture (4200 BC to around 2000 BC) and Corded Ware culture (2900–2450/2350 cal. BC). All these cultures are important in shaping the genetics of modern Finns. By contrast, Butovo Culture (8000-6000 BC) was an eastern Mesolithic microblade culture and Karelian Oleni Ostrov culture was close to it or part of it (and now we have heard that they had a high amount of WHG). Butovo Culture probably shaped more the genetics of Finns compared to Balts. FYI, at the moment we do not know the y line of people who brought to Finland the small Arctic element which is concentrated in the north.

Unknown said...

Kristina
You're of course right
But that kind of nuance is a bit unfashionable here
Nothing but proof of sweeping migrations from the east will suffice , for some

postneo said...

Oops sorry for the double post.

@davidski
saw the map. I am not talking about the fertile crescent but high levels of ANE east of the persian gulf, baluchistan(not there in your map), afghanistan.

@maju
caught up with your response. I still do not see anything diagnostic of IE. social stratification, warfare, endogamy are a function of population density. Egyptians built massive kurgans and were highly socially stratified.

I don't subscribe to romantic notions of patriarchal kurgan baddies and earth mother worshipping goodies. Only shifting strategies for survival. Small populations cant afford war fare. neighbours are vital to avoid inbreeding and barter goods. Ritual warfare/bitching/tall tales is all you see among hunter gatherers.

Davidski said...

How is Baluchistan relevant when the Neolithic expansion started in the Fertile Crescent and there was plenty of ANE on the Eurasian steppe, and there still is in fact, despite the recent influx of East Asian admixture?

Unless there's something South Asian in Europe that can't be attributed to Roma migrations, then all of the ingredients for the Yamnaya are located in the Near East and on the western steppe.

Nirjhar007 said...

@David
''Unless there's something South Asian in Europe that can't be attributed to Roma migrations, then all of the ingredients for the Yamnaya are located in the Near East and on the western steppe.''
K=8 Results will be fascinating! anyway Since 7 male y-dna will be published i bet R1a will be there in at least 4 Samples 2 will have I and the last one no idea:)!
Consider that as my prophecy;)....

Marnie said...

@Davidski

I looked at your map.

I also had a look at the Anzick paper, which has an admixture run with an ANE like component that matches the hotspots on your map.

If I am right, the ANE component appears that it could have dispersed into Europe from four points: 1. Finland, 2. Poland, 3. Turkey and 4. Caucasus.

In order to understand the timing an dispersal of "ANE", as you define it on your map, it would be necessary to do a lot of sampling across time and space to pinpoint flow from the regions mentioned above.

Do you have the Kalash in your data set?

Marnie said...

@Davidski

Never mind about my Kalash question. I see you've already covered this on previous posts.

Maju said...

"H+U5+U4+U2=~60% for typical Europeans. Even if you think all H is of Mesolithic origin, that's unproportional for many Europeans".

Well, now you're talking serious stuff but also a bit of splitting hairs when you look at it in depth. (You should add lineages like U8(xK) and HV0 but that's less important).

This problem appears already in EEF themselves, said in Lazaridis et al. to be more similar to Lochsbour than to West Asians, yet with an appearance of having much more West Asian haploid lineages than native European ones. This is a founder effect in Thessaly (or call it "drift" if you wish) that affected the whole process of Neolithic colonization further West and North.

But even if we accept this as a given and consider EEF-ness rather than West-Asianness (more elusive for lack of ancient data) we still need to double the figures of mtDNA H (particularly) to make your average modern European genetic pool. This implies a second "mystery" population with very high H frequencies (because of another founder effect, say in ancient Portugal or Brittany or wherever).

The little evidence we have for such a H-rich second population points to Neolithic Portugal and Germany's true Bell Beaker sites (excluding the more than dubious Kromsdorf case). These are just two data points and no other data points follow this pattern so far. Possible related sites of the Megalithic arch like Gokhem or Basque sites are not so extremely high in H - although they are definitely modern-like in this aspect and therefore vaguely supportive of an Atlantic origin of this excess H (no other possible sources like Kurgan sites have that key pattern of greater H, rather the opposite).

And that's what we have so far. Several studies (both genetic and archaeological) suggest a re-aboriginalization of sorts from the Atlantic and/or from "invisible" scattered populations around the more visible EEF neolithic sites. And I'm indeed in this line of thought. The fact that archaeo-statics show much greater demographic impact with the Atlantic Neolithic/Megalithism than with Bell Beaker or Corded Ware, make me think that this is the key source of this excess H. Excess U (less impressive) may come from this or other sources, unsure.

...

Maju said...

...

"If H did exist in Mesolithic Europe without sometype of near eastern-intrusion(which I think it very well may have), it was very very rare".

It was reasonably common at least in Iberia, particularly towards the West (there's clear evidence of that now). But even if you want to imagine H as arriving with EEFs, there must have been a place or several where there was a founder effect (say "drift" if you wish) that magnified its frequency to very high values, while minimizing the other Neolithic lineages. The ultimate origin of H is not really that important per se, what matters is that its current and paleohistorical frequencies show a process of de-EEF-ization and re-aboriginalization, also apparent in autosomal DNA. Now if we see a re-aboriginalization it's likely that the lineage is aboriginal (what is backed by some aDNA data) but that's just a less important aspect of the problem, hope you understand.

"Maybe it was very popular in Iberia, I don't know, but that would be strange considering their close relation to Mesolithic people who lived north of them".

I agree that it's a bit of strange on first look but notice please that we have a huge blank of data in France (where genetic research is legally difficult it seems), so the North of Iberia in aDNA is already Germany and the Benelux and that's quite a distance. Only one French Megalithic site (n=3) has been sequences and no, absolutely no, Paleolithic DNA whatsoever comes from that state.

Also it's possible that Iberia (particularly Portugal) played a major role in that Atlantic backflow. Both Megalithism and Bell Beaker originate precisely there, an area that also holds the first civilizations of Atlantic Europe, so...

postneo said...

@davidski
How is Baluchistan relevant when the Neolithic expansion started in the Fertile Crescent and there was plenty of ANE on the Eurasian steppe, and there still is in fact, despite the recent influx of East Asian admixture?

Baluchistan is relevant as a relic and a proxy for old iranian. My point is that ANE did not arrive from the steppe to west asia as is sometimes being suggested but was already there in the neolithic in Iran and Baluchistan. In terms of y lineages there was R1 and R2 in Iran and south asia.

Davidski said...

How do you know there was R1 in Iran during the Neolithic?

Maju said...

@Mike: In most regions there's no extreme collapse that could explain such "drift" and anyhow, and this is important, it is effectively impossible that drift leads to the same results everywhere: if you have ten jars with assorted red, blue and yellow marbles and proceed to simulate the drift process by taking random halves of each in successive steps, you won't end (unless you're extremely lucky, to the point of being a lottery winner) with ten jars of red marbles, but rather with a different result in each jar.

So the process, if it must be attributed to drift or similar bottleneck process (founder effect for example), must have happened in a single "jar" and then spread all around from that origin.

This is fundamental. Another aspect is that we see associated to that H increase a partial re-aboriginalization of the autosomal genetic pool. This is the other aspect which affects to whether H (and possibly R1b) is immigrant or aboriginal, because it'd be most strange that an increase of aboriginal (WHG) autosomal DNA was associated to both haploid lineages that are not aboriginal.

This and other reasons (confirmed presence of H in Paleolithic SW Europe, even in large frequencies probably, West-centered structure of European R1b) make me think that mtDNA and Y-DNA R1b-M412 re-expanded in Chalcolithic from European sources, with some EEF autosomal genetics but also a notable greater apportion of WHG than what EEFs had.

But this is the other aspect of the problem: even if H and R1b would be immigrants, their expansion (at the very least that of H) is clearly associated to a major increase of WHG autosomal DNA, no doubt.

postneo said...

@Davidski said...
How do you know there was R1 in Iran during the Neolithic?

Its a projection because there is both R2 and R1(r1a, r1b) today in iran and india. R2 in central asia and europe is derived unless we get better evidence.

Balaji said...

The Middle East and Europe both got an infusion of ANE some time in the Copper Age. It is likely that the ANE ultimately came from the same source. Also it appears that whoever brought ANE to Europe and the Middle East also brought along some ENA. I say this because of f3 statistics that Davidski published.

Middle East
f3(Assyrian;Dai,Sardinian) -0.00059462 0.000194433 -3.05822
f3(Ashkenazi;Dai,Sardinian) -0.000297346 0.000155065 -1.91756
f3(Armenian;Dai,Sardinian) -0.000435056 0.000143102 -3.04019
f3(Cyprian;Dai,Sardinian) -0.000721083 0.000160358 -4.49671
f3(Druze;Dai,Sardinian) -3.8934E-005 0.00017135 -0.227219
f3(Sephardic_Jewish;Dai,Sardinian) -0.00134829 0.000154977 -8.69989

Europe
f3(Spanish;Dai,French_Basque) -0.000975007 0.000124464 -7.83365
f3(Greek;Dai,French_Basque) -0.000365561 0.000151607 -2.41124
f3(East_Sicilian;Dai,French_Basque) -0.000569748 0.000184245 -3.09234
f3(Tuscan;Dai,French_Basque) -0.000432676 0.000151683 -2.8525
f3(Portuguese;Dai,French_Basque) -0.000915428 0.000242182 -3.77992
f3(Bulgarian;Dai,French_Basque -0.00118302) 0.000144746 -8.17306
f3(Romanian;Dai,French_Basque) -0.00125626 0.000141042 -8.90695
f3(North_Italian;Dai,French_Basque) -0.0002888 0.000159868 -1.80649
f3(West_Sicilian;Dai,French_Basque) -0.000558259 0.000196153 -2.84604

Chad said...

Oh my... Where to begin?!? Maju, you went off the deep end a long time ago and just keep digging yourself deeper into a hole from which you wont return. If you wish to have anyone take you seriously after this paper, it might be best to discontinue that type of talk.

WHG didn't re-expand from Iberia. That is completely laughable. R1b is not expanding out of Iberia and it certainly isn't 30kyo.

The resurgence of WHG in Germany is going to be thanks to Blatterhohle, Ostorf, and related HG groups that were surrounding Neolithic settlements.

Krefter said...

A problem with R being EHG, is there are no basal R lineages in eastern and northern Europe, except maybe forms of R1a. EHG, SHG, and hunter gatherers from central-east Europe defintly have many descendants today, and so we should expect to see basal R.

This is a wild guess, because I haven't looked alot at modern data. ANE paternal ancestry north of the Caucasus is represented by Q and ANE paternal ancestry from south of the Caucasus is represented by R.

I've heard of basal Q-types in Europe, but I don't remember very much about it.

Besides P, C, and I, N looks like the best canidate for EHG Y DNA results. But I don't whether or not N in Europe is very diverse or not.

Nirjhar007 said...

@Postneo
We can bet on U2-Mtdna also to bring ANE in SC ASIA from paleolithic times at least.

Krefter said...

"WHG didn't re-expand from Iberia. That is completely laughable. R1b is not expanding out of Iberia and it certainly isn't 30kyo. "

WHG could have risen because of almost any reason in Neolithic western Europe, because pretty much everyone was WHG-ENF.

All we can say is WHG did rise, and we can't know exactly who caused that rise. His theory isn't totally laughable because he knows about archaeology.

In recent years people have become obsessed with hunter gatherer vs farmer. These are realistic categories genetically, but as you go 100s of years into the Neolithic it probably isn't that simple. Hunter gatherer tribes could have taken up farming, and admixed with high-ENF pops, and created Gok2.

pnuadha said...

Krefter

How is that not relevant? You can't assume Poles and Germans have similar ANE, when they obviously are very different genetically.

It is not relevant! There is no other way for me to say that. Look at how the argument is structured. It goes as follows.

If West/Northwest Europeans ultimately got their ANE from Corded Ware then it was mediated by German and Polish CW. Modern Germans and Poles certainly can be said to have more German/Polish CW ancestry than Northwest Europe. So if CW was the dominant source for ANE in West/Northwest Europe then West/Northwest Europe should have a lot less ANE than modern Poles and Germans. But they don't! 16% seems to be an upper bound for the ANE in both the Poles and Germans - so what if the Germans have less; ITS THE UPPER BOUND OF THE TWO POPULATIONS THAT IM USING. So the Germans and Poles don't really have any more ANE than West/Northwest Europeans (French, English, Scottish, and Orcadians) despite having more CW heritage. Hence, we need an alternative source for ANE in West/Northwest Europe independent of the CW.

I don't want to get bogged down by meaningless discrepancies when it does change the validity of the argument.

Anonymous said...

@postneo
"We cant expect modern day armenians, iranians and kurds to have the same ANE levels as before since they have been increasingly influenced by levantine populations.. genocide even. If you go east of Iran to more sheltered areas what happens to ANE % ?"

the genetic impact of arabs in iran and kurdistan is absolutely overrated and modern persians or kurds are very similar to pre-islamic kurds and persians. They have not so much ANE because it was never important there unlike in the caucasus, hindukush and siberia. Persians and kurds are not dominated by Q or R which can be associated with ANE and cluster rather with near eastern populations of non-iranic origin than with other iranic people in central asia. The genetic impact of indo-iranians in iran and kurdistan was smaller than in central asia and even north india. The deserts in central iran acted as barrier for population movements and today ANE and R1a is much higher in eastern iran than in western iran, what also disproves that ANE and r1a orginated in the near east.

Krefter said...

Davidski, do you think you can make a composite of Motala12 and a Caucasus group with higher ANE than Armenians to create a Samara Yamna zombie? Then see if north Euros fit as a mix of that zombie and Gok2? Or even better see if Lithuanians fit as a mix of that pop and Loschbour or Motala12?

Like what you did with this PIE zombie.

http://eurogenes.blogspot.com/2014/08/guessing-game.html

I don't if anyone here is very knowledgeable about Y DNA N, but I looked at Eupedia's page for it, and it seems to be pretty diverse in Europe.

http://www.eupedia.com/europe/Haplogroup_N1c_Y-DNA.shtml

It could be EHG. So, then Balts could be a mix of Yamna and something SHG/WHG-like.

Davidski said...

It'd be easier to create a mix like this with the K8 proportions. The outcome would also be more precise. And I could easily plot it on the PCA too.

Krefter said...

"It'd be easier to create a mix like this with the K8 proportions. The outcome would also be more precise. And I could easily plot it on the PCA too."

I think this will be an interesting visual.

Yamna: 25 ANE, 35 WHG, 40 ENF.
BHG: 20 ANE, 80 WHG
Funnel Beaker: 50 WHG, 50 ENF.

If you can, can you make a triangle of those three?

Yamna: 25 ANE, 35 WHG, 40 ENF.
BHG: 20 ANE, 80 WHG
Greek Neolithic: 20 WHG, 80 ENF.

I triangle for this would probably have all Europeans in it.

Unknown said...

@ Skilur

"The deserts in central iran acted as barrier for population movements and today ANE "

Yes but prime landscape for founder effects.

@ Maju

I was not arguing pro or con the expansion of mtDNA specifically. I havent really followed mtDNA enough. Alls I was pointing out was the real-time situation on the earth. We don't know exactly how much of a population drop is needed for drift to take effect, but it must be massive. based on studies from 18th century African buffalo plague, it was estimated that some 85% of population needs to drop for drift to 'effect'.

BUT - if you combine drift, selection, and a bit of admixture, then the the latter need not be so great. Many forces work on DNA structure, not just pop A meets pop B.

Helgenes50 said...

David,

If Yamnaya is a mix of West Asian and Eastern Europeans
Knowing that the West Asian contrary to the ANE was not present among the West Europeans, not before the bronze age why don't use it to evaluate the IE ratio
The Sardinians, the Basques and the SW French lack of it and that is not the case for the ANE.
The West Asian decreases from East to West, what perfectly fits with the IE advance in Europe

Shaikorth said...

The typical "West Asian" component appearing in ADMIXTURE runs has a north/south cline too, which is actually stronger than the West/East one. Poles don't usually have more than English. Basques and Sardinians in the southwest are unique cases.

Lack of West Asian in Basques, Finns and Sardinians might support the component having an IE link, but it can't be an universal IE marker if we assume a steppe expansion because of where it peaks among IE-Speaking Europeans (South Italy and Greece)

http://1.bp.blogspot.com/-LIug8kEPKW4/UFtC9XP-VsI/AAAAAAAAD1A/HTmsEscIGoI/s1600/MDLPwestasian.jpg

Ryukendo K said...

Well, we're speculating without an ounce of extra information here. Why don't we try to figure out ways to analyse the data in preparation?

@ Davidski
You mentioned that you did not discover substructure in ANE, because all the ANE variation that you seived out from Amerindians, kalash, Caucasus, E.Euro and etc. placed in the same position in PCA.

I expect you did the PCA with other West Eurasian populations? If so, it is quite expected now that the ANE zombies would score in the same position, because this is an analogous situation as the ANE point of the 'triangle' we were looking over.

However, did you try a PCA containing only the ANE zombies from each population and nothing else? If you did, PCA would force the variation within ANE from different pops to be revealed, no?

Just an idea. Could we sieve out ANE from a South Indian pop, Baloch, Kalash, a N.Cauc pop, Kazakh, Tajik, Amerindian, a Ket, Lithuanian, Mari and Motala?

Later we can add the sieved ANE from EHG and Yamnaya.

Let there be light!

Of course, differences in data quality might ruin the whole thing.

Helgenes50 said...

@Shaikorth

Yes indeed, in South Europe like Italy and Greece, the West Asian is not a good IE marker, because of recent migration, but in west Europe, the history is maybe different, most of it is certainly old, from the metal age
I don't believe in a Roman origin, like in NW France nor in Britain and certainly not in Ireland

Nirjhar007 said...

@Guys
What about the South- West Asian Component? How that scores around Europe?according to ''the Genographic Project, British (all UK) have 50% Northern European (identified with the hunter-gatherers), 33% Mediterranean (identified with farmers), and 17% Southwest Asian, so defined: "This component is found at highest frequencies in India and neighboring populations, including Tajikistan and Iran in our reference dataset. It is also found at lower frequencies in Europe and North Africa. As with the Mediterranean component, it was likely spread during the Neolithic expansion, perhaps from the eastern part of the Fertile Crescent." So, it harmonizes well with Zagros-Zarzian PIE model. The difference of the ANE component and the Genographic SW Asian is that the ANE seems to include the Genographic Native American and SW Asian components, while if they are distinguished they do not overlap. Native Am. component in Asia is given only at 4% in Altai (Siberia) and Mongolia, showing the probable region of origin of this component, evidently later submerged by others.
SW Asian is 58% in Western and Southern India, 44% in Pamir, 42% in Iran. The fact that India has more than Iran is because India was less touched by the Mediterranean component, and the Neolithic expansion was dominated by the SW Asian component.
The Genographic in Eastern Europe gives e.g. the percentage of Russians: 51% North European, 25% Mediterranean, 18% SW Asian, 4% NE Asian. Tatars of Russia (settled on the Volga) are also interesting: 40% N Eur, 21 % Med, 21% SW Asian, 16% NE Asian. Siberian people from the Altai have 53% NE Asian, 22% SW Asian, 17% N Eur, only 2% Med, 4% Native American.
Also interesting Northern Caucasian (including Daghestani and Abkhazians): 46% Med, 33% SW Asian (close to the ANE percentage in Lezgins of Daghestan), 16% N Eur, 4% NE Asian.
It is important that the Mal'ta boy, foundation of the ANE, was R* for Y Dna, and the R Hg is present also in America as R1-M173.''???

postneo said...

@nirjhar
good point about possible U2 and ANE correlation

@Skilur
the genetic impact of arabs in iran and kurdistan is absolutely overrated and modern persians or kurds are very similar to pre-islamic kurds and persians.

I agree but the impact from early farmers would be important. Iran is an interesting bridge you have G,I,J, IJ* in the west and R in the east.

@ryu
Just an idea. Could we sieve out ANE from a South Indian pop, Baloch, Kalash, a N.Cauc pop, Kazakh, Tajik, Amerindian, a Ket, Lithuanian, Mari and Motala?

On zombies. If available, the more extreme spread populations are the best candidates. Look at Bhutan. similarly Sri Lanka, Yemen, Assam should be good candidates.

Helgenes50 said...

The west Asian is not supposed to be the totality of the IE Ancestry, but just one of the component and from the results of the ancients genomes, what is sure right now this one is not present among the Neolithics, in Europe of course, I am speaking of the K13 West Asian, the different components don't having the same meaning in all the calculators

If the Yamnaya is supposed to be a mix of EHG and Armenian-like, the West Asian is not negligible

Chad said...

Krefter,
Clearly, if he thinks Baalberg is Kurgan, he is stuck in the 60's. The farthest the eastern influences in Baalberg go are Lengyel and Baden. Neither one is Kurgan and neither have ANE. This is clear.

Nirjhar007 said...

Can you give a K link which also has SW Asian as Run?

Shaikorth said...

Helgenes50,

Eurogenes K13 splits the "Armenian-like" stuff into West Asian and East Med. If we take one of the ancient genomes as an example, BR2 has 5.89% of MDLP-22 West Asian component (frequency map in my previous post) but less than 1% Eurogenes K13 West Asian. It has about 10% of K13 East Med though. Eurogenes Hunter-Gatherer test has just one West Asian component (Anatolian Farmer) and there BR2 reaches 5% like in MDLP K-22.

Perhaps Yamnaya's K13 West Asian/East Med ratio turns out useful in tracking their specific West Asian signature. But things rarely turn out that clear-cut.

Maju said...

@Chad: your insults aside (I'm trying to stay cool in spite of all), if it's not Iberia it will be France or Britain or whatever. The exact location of the origin that re-expansion is not the key point, the key point is that the semi-aboriginal re-expansion happened and did it almost certainly within Megalithism (plus Bell Beaker?) The origin could be specific or diffuse, could be in Portugal or Scotland... that's a second issue to debate, not the key point.

Maju said...

@Krefter: "WHG could have risen because of almost any reason in Neolithic western Europe, because pretty much everyone was WHG-ENF.

All we can say is WHG did rise, and we can't know exactly who caused that rise. His theory isn't totally laughable because he knows about archaeology".

Thank you very much. I truly appreciate your open-mindedness.

"In recent years people have become obsessed with hunter gatherer vs farmer. These are realistic categories genetically, but as you go 100s of years into the Neolithic it probably isn't that simple. Hunter gatherer tribes could have taken up farming, and admixed with high-ENF pops, and created Gok2".

A very reasonable synthesis, yes.

A modern comparison (with due caution, but still illustrative) could well be the ongoing expansion of Mexicans and other "Hispanics" in the USA, who clearly carry a big share of Native American genetics but are not quite Native Americans (yet not European creoles either) and whose Native ancestry does not even directly come from the US territory as such. Take the comparison with a pinch of salt but it does make some sense, right?

Maju said...

"I don't if anyone here is very knowledgeable about Y DNA N"...

Kristiina, who commented above and is hopefully still following the discussion, knows A LOT about these Uralic-related genetic matters, not just Y-DNA but also mtDNA. We had some very interesting discussions back in the day at my nowadays neglected blog.

What is clear is that N1 comes from East Asia and that this migration happened via Siberia (the taiga by Holocene parameters) after the LGM, mirroring in reverse the older "paleo-Siberian" (ANE-like, Q1) one towards Beringia and America. It's also clear that some of those East Asian genetics (mtDNA at the very least: C*, C1 and C4) were already in Eastern Europe (Ukraine, Karelia) very soon in the Holocene.

This antiquity of "Uralic" genetics in Europe allows for a lot of scatter later on, particularly if they hijacked migrations of Dniepr-Don/Pitted Ware first and Kurgans later, let alone later migrations (Vikings for example).

Notice that at least part of the gray area north of that map is largely occupied by its brother lineage N1b.

Hope this helps.

Maju said...

@Mike "I havent really followed mtDNA enough".

That's a pity. For starters you may want to take a look at these entries:

→ http://forwhattheywereweare.blogspot.com/2013/09/basque-and-other-european-origins.html

→ http://forwhattheywereweare.blogspot.com/2014/08/chalcolithic-mtdna-from-atapuerca-still.html

The first one is my own synthesis of many works, focusing on the three best sampled regions (Germany, Basque Country and Portugal), the second one is on a particular scholarly paper (Gómez-Sánchez, Olalde et al.) showing similar results. They show more varied samples but a criticism I make is that they pool Kromsdorf with "true" Bell Beaker from Germany, what dilutes the dramatically high H figure of the later. Otherwise it's fine.

Re. "drift", in this case what we are actually looking at is founder effects instead because of the example of the jars and marbles. You need first a single jar where that "drift" happens and then an expansion from that jar. That's rather said to be a "founder effect" - and it has quite different implications about what kind of process we are talking about.

In the appeal to drift, we are facing a situation in which all the jars "magically" converge to very similar results, what is effectively impossible. Instead in the founder effect scenario you have a single original jar from where the marbles, already with a defined pool, spread around.

You can still appeal to positive selection, but that would be a third different scenario. So far nobody has been able to prove such selection, so I tend to disdain it. Also it's not just lineages but autosomal DNA, so what is being "selected for" are rather whole genomes and not a particular mitochondrial type. This fits better with the notion of backflow from (I understand) the Atlantic fringe (whose peculiar Neolithic cultures and dynamics have been largely ignored so far but no doubt had an impact at least in Megalithic times).

Shaikorth said...

"Kristiina, who commented above and is hopefully still following the discussion, knows A LOT about these Uralic-related genetic matters, not just Y-DNA but also mtDNA. We had some very interesting discussions back in the day at my nowadays neglected blog.

What is clear is that N1 comes from East Asia and that this migration happened via Siberia (the taiga by Holocene parameters) after the LGM, mirroring in reverse the older "paleo-Siberian" (ANE-like, Q1) one towards Beringia and America. It's also clear that some of those East Asian genetics (mtDNA at the very least: C*, C1 and C4) were already in Eastern Europe (Ukraine, Karelia) very soon in the Holocene. "

Those old STR-based studies should be revised at some point in light of SNP results that we have from new sequencing. SE-Asian N is an early diverged branch, but Europe has even earlier clade (the divergence time for P189.2 is long before LGM) that is absent in Siberia and East Asia. The history of N needs to be confirmed with ancient DNA.

The Hong Shi et al. map also shows movement of N through Northern Eurasia through Yakutia and Northern Siberia. This is almost certainly wrong, and it happened through Southern Siberia and the Asian steppe, with Northern Siberia acting as dump for small population movements. Adamov's (RJGG 2014) research shows Yakutian N is just a bottlenecked offshoot. Yakuts are known to be very recent migrants to that area to boot. Nganassans, Nenets and other Siberian forest zone N will almost certainly turn out to be a similar case.

This is the up-to-date N tree.

http://i1303.photobucket.com/albums/ag150/Alex_Chartorisky/SNP-N-TREE5_zps6e68bd8b.jpg

Chad said...

Bell Beaker mtDNA

Quedlinburg
1-U5
1-H5
1-T2
1-H13

Rothenschirmbach
1-K
1-H5
1-H
1-H3

Quedlinburg
1-H
1-U5
1-J
1-H4

Benzingerode-Heimburg
1-H1
2-U5
1-W
1-T2

Alberstedt
1-H
1-H3

Eulau
1-H
1-W
1-T1a

Karsdorf
2-U4
1-U5

Damsbo
1-U4
1-U5

Maju,
I am not sure how you deduce that true Beaker is 80% H, when hardly anything fits that bill. Several of those H's aren't of the H1 and H3, but local variety.

Marnie said...

@ryukendo kendow

"You mentioned that you did not discover substructure in ANE, because all the ANE variation that you seived out from Amerindians, kalash, Caucasus, E.Euro and etc. placed in the same position in PCA."

"I expect you did the PCA with other West Eurasian populations? If so, it is quite expected now that the ANE zombies would score in the same position, because this is an analogous situation as the ANE point of the 'triangle' we were looking over."

The Kalash, Ket, and Amerindian groups cleanly separate for all components in the Admixture run in this paper (supplementary datat):

http://www.ncbi.nlm.nih.gov/pubmed/24522598

So I doubt that the inferred "ANE" like components in these population could have come from the same source in the last 12,000 years.

It's much more likely that these populations share low level ancestry from an archaic population.

Keep in ming that the separation between the Karitiana and the Kalash had to have occurred more than 15,000 years ago. (ie. due to Beringia flood dates, time to get to the Amazon).

Maju said...

@Saikorth: Very curious that upstream branch in the Balcans, yes. Still it doesn't seem to disavow the previous work because it can well be seen as yet another East-related erratic as the C1 in La Braña (and few modern Europeans), which would have migrated westwards with P1 (ancestor of both Q and R, as well as of some P1* to be found in India, particularly towards Bengal), surely around the UP genesis.

Krefter said...

Davidski, I'm moving this post down, in-case you can't see it.

"It'd be easier to create a mix like this with the K8 proportions. The outcome would also be more precise. And I could easily plot it on the PCA too."

This is an edit of my previous post.

Yamna: 25 ANE, 35 WHG, 40 ENF.
BHG: 20 ANE, 80 WHG
Late EEF: 44 WHG, 56 ENF.

IMO, this is a very realistic description of the three main meta-populations that inhabited Europe ~5,000YBP, and admixed to explain most modern Euros ancestry. Ancient DNA has already confirmed all of them, except maybe Yamna.

I'd like the next one to test whether Balts can even be modeled as Yamna+something with ENF.

Yamna: 25 ANE, 35 WHG, 40 ENF
#1 EEF: 50 WHG, 50 ENF
#2 EEF: 70 WHG 30 ENF
#3 EEF: 60 WHG, 10 ANE, 30 ENF

The next set

Yamna: 25 ANE, 35 WHG, 40 ENF
SHG: 16 ANE, 84 WHG
EHG 35 ANE, 65 WHG

Maju said...

@Chad: "I am not sure how you deduce that true Beaker is 80% H".

→ http://forwhattheywereweare.blogspot.com/2013/04/central-european-bell-beaker-mtdna-88-h.html

Quoting (BB sites):

Quedlinburg XII: 3 H-CRS (H1?), 1 J
Rothenschirmbach: 2 H3, 1 H5
Alberstedt: 1 H-CRS (H1?)
Total (simplified): 7 H, 1 J


That's 88% H, 12% J. By contrast data from the same study for Unetice has 0% H, although it has a sizable apportion of U:

Total (simplified): 4 U5, 3 I, 3 T2, 1 T1, 1 U2, 1 U*, 1 W, 1 X.

(Also interestingly I, which can be seen as associated to Kurgan expansion as well).

I'm not sure where your samples come from, sorry (no references) but, assuming they are correct, it indicates that there are some sites (Benzingerode-Heimburg, Eulau, Karsdorf and Damsbo) that do not fit the pattern first found by Kristina Adler and therefore that BB in Germany were less homogeneous than suggested in that study. There are still many sites with very high frequencies of H and that is still interesting and probably points to underlying differences from previous periods.

I'll see if I can figure out a geographic or some other pattern that makes sense but anyhow, I'd appreciate your reference, as it could facilitate my research and confirm your data as valid.

Shaikorth said...

"Very curious that upstream branch in the Balcans, yes. Still it doesn't seem to disavow the previous work because it can well be seen as yet another East-related erratic as the C1 in La Braña (and few modern Europeans), which would have migrated westwards with P1 (ancestor of both Q and R, as well as of some P1* to be found in India, particularly towards Bengal), surely around the UP genesis."

Is there really a reason to link early West Eurasian C with P1? We see C already in Kostenki14. The most western paleolithic K derivative we know of is Ust-Ishim, which is closer to NO than P.

Chad said...

You're missing a lot of new mtDNA samples. H is down around 50% now.

Chad said...

Only 11/23 are H, outside of Kromsdorf now.

Quedlinburg is 50% 4/8

Davidski said...

rk,

I haven't tried extracting ANE from different regions separately. But the very first ANE cluster was based on West Asian, Central Asian and Amerindian samples, and it was identical to the K8 ANE, which is based on European and Siberian samples as well.

Krefter,

Here are those plots...

https://drive.google.com/file/d/0B8XSV9HEoqpFdkdpUERjZU5uMUE/view?usp=sharing

https://drive.google.com/file/d/0B8XSV9HEoqpFam45NDluWEZERjA/view?usp=sharing

https://drive.google.com/file/d/0B8XSV9HEoqpFYjBmU094QmRYeWM/view?usp=sharing

Matt said...

Ryu: "@ Davidski
You mentioned that you did not discover substructure in ANE, because all the ANE variation that you seived out from Amerindians, kalash, Caucasus, E.Euro and etc. placed in the same position in PCA.


Ryu, are you sure about this?
There is specific software to allow us to plot the ANE segments of samples together on a PCA. This is the publicly available PCAdmix software used by Moreno-Estrada et al to reconstruct the European genotypes in the Caribbean - http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3828151/ which showed the resulting European genotypes to be between Sardinians and Basques (and also showed the differing structure of Native American genotypes).

However, I think the last time I mentioned this to David, his comment was that it wasn't technically possible to run anything like this with the ANE segments, which seems fair enough.

But I'd be surprised if there was no structure to ANE segments between populations, as Mal'ta is from 24000 years ago! And personally I doubt there was much genetic contact between MA-1's indirect descendants in East Siberia and West Eurasia (I don't think there was this mammoth steppe conveyor belt, not really, that's just my opinion though).

Laziridis's tree models show significant divergences should exist between the ANE which contribute to Europeans and Native Americans, if only because of high drift between populations after ANE admixed into them. The drift in Laz's model between Eurasian ANE-Amerindian ANE is the half again as between WHG and ANE and that is before layering in additional drift after admixture of Amerindian ANE and East Asian.

Drift on the ANE portions alone in Laziridis's model does not seem to contribute strongly, however, independently. That may be because strong drift happens only under certain conditions of reduced population size, so it may be that drift is not perhaps as informative of the time depths populations have been in contact as I might expect, and populations separated for thousands of years may experience little drift. Maybe the Western ANE just didn't experience much drift, as they got all their drift early on when separating then stayed at relatively high population sizes, which would surprise me but could be true. That's obviously concerning if we're trying to treat drift as a neutral "clock" of relatedness through time.

Shaikorth said...

I agree, and it would be quite interesting to see ANE's extracted solely from South Asian, West Asian and European samples respectively. Perhaps even one from Siberians.

Chad said...

Correction here, Maju.
Here is German Beaker, outside of Kromsdorf

H-12/26 only 3 are H1or3
U-7/26
T-3/26
W-2/26
K-1/26
J-1/26

Most H's look local or Danubian derived from some point, and the rate is not much more than prior cultures in the region.
You can either look at Ancestral Journeys, Krefter's page, or the papers by Brandt and Brotherton, from 2013.

Maju said...

@Chad: "You're missing a lot of new mtDNA samples".

If you notice I haven't blogged since August and mostly I haven't checked the RSS feed either. So no surprise. Still asking for references, which you fail to provide, why?

Chad said...

Also,
Baalberg (3800-3400BCE) is more of a TRB/Lengyel/Early Baden mix. We saw from those results that neither Lengyel, nor Baden(prior to 2800BCE) had ANE. The abstract does state that no ANE was found prior to 3000BCE, which makes perfect sense. If Baalberg ever had ANE, it would've been present in the surrounding group samples prior to 3000BCE, as well.

Maju said...

@Shaikorth: "Is there really a reason to link early West Eurasian C with P1?"

Well, K-M526 seems to have originated in what is now Indonesia or Malaysia and P1 (as NO and other sublineages) is an offshoot, with the peculiarity that it seems the bulk of what migrated westwards through North India (surely just before the earliest UP). I understand that this migration westwards is associated to that of N and R (looking at both sides of the haploid data instead of just one can give you depth of vision, see HERE for my best synthesis so far).

I don't see any reason to imagine more than one main migration westwards from the earliest Eurasian homeland of H. sapiens in Tropical and Subtropical Asia (even if it may have got some nuances), so I conjecture that all those oddball lineages with some apparently ancient connection with East Asia (a very upstream branch of N, an odd C1 sublineage) arrived in the same process.

I may be wrong, naturally, but that's my logic in any case. More so when we see those lineages not in or near Siberia but towards the Mediterranean.

Chad said...

I gave the references. Look above.

Ancient central European mtDNA across time (Brandt, Haak et al. and Bollongino et al.) Brotherton has stuff as well. Look at Krefter's page or Ancestral Journeys.

Unknown said...

Shaikorth
Thanks for sharing those image links
They nicely show the structure if N1.
But please define what u mean by "Asian steppe". That's a rather non-specific entity

Maju said...

Now you gave me the reference (not before), thank you, Chad.

Something I notice is that all the samples (these and the other ones) come from the same region of the Saale, also very sampled for other periods and very interesting no doubt. It is a key region in Kurgan development (if we accept the Baalberge → CW model) but it is clearly outside the Megalithic area and hence maybe not so informative for what happened West of it.

In any case the data you give seems to de-hype the BB association with excess mtDNA H, what is something I do welcome (I was surely relying too much on Alder's model and other Bell Beaker fans, when I've been all the time proposing that Megalithism seems more important instead, with BB being just a late epiphenomenon).

It is still true that BB samples on average (at least per Brandt figures) are the only ones with more H than modern Central Europeans (48% vs modern 41%).

I'll await more data from more westerly areas anyhow, we can't extrapolate to all or much of Europe from a small region of Germany alone (that's as good or rather as bad as extrapolating only from the Basque Country, really). We do need more data and preferably from Western Europe.

Unknown said...

Yes that Balkan N1 is weird
What actual population was sampled. ?

Shaikorth said...

That N type can be found outside Balkans in East-Central Europe. In modern times C-V20 is very rare but found around Europe. La Braña and the neolithic Hungarians were not the only C's either, UP Kostenki14 was from Middle Don region. I'd wait for P of similar age to ancient C's to pop up in these regions before considering attributing this to one migration of P and stragglers.

Chad said...

I actually did give that info in a previous post that you may have missed.

We do have a ton of mtDNA samples from Western Europe. France doesn't have any great amount of H or U.

Megalith France, 4200BCE
1-N1a
1-U5b
1-X2

Treille 3000BCE
6-U
6-J
6-H
4-X
2-T
2-K
2-HVO
1-V

Megalith France 2750-2725BCE
2-K

Unknown said...

Shaikorth
Of truly a basal N, then who knows what other basal lineages would pop up in the Balkans ?
Possibly all later found through rest of europe

Ryukendo K said...

@ Davidski @ matt

Thanks for the answers.

Hmm, what is the technical issue exactly though?

The previous time you tried adding excess ANE segments into a Lezgin genome to produce the PIE 'Frankenstein' which you then plotted in PCA. Surely it means that ANE segments themselves can be manipulated into a 'Frankenstein' and made to plot in PCA?

In a West Eurasian PCA, variation in proportions of ANE:WHG:EEF dominate all the dimensions, but if we tried this on a PCA with no other samples other than the ANE zombies from different populations, my understanding is that PCA would have to find dimensions of variation, even if they are small, no?

Hopefully, this can help to discern purely ANE-mediated flows from Europe/Siberia to West Asia or vice versa.

If this method works, it can also be applied to WHG in Europe, and Danubian vs. West Asian vs. Mediterranean EEF.

Krefter said...

I figured out how to plot fake ANE K8 scores on an ANE K8 PCA. It's very useful.

"Using the "fateful traingle" to discover who Euro's non-Yamna ancestors were".

http://www.anthrogenica.com/showthread.php?3747-Using-the-quot-fateful-traingle-quot-to-discover-who-Euro-s-non-Yamna-ancestors-were

Chad said...

That's what I do, but stats can narrow it down.

Maju said...

@Chad: Thank you for the info. I knew about at least one of those Megalithic samples and of course Treilles (not sure about the other). It's still a very bad coverage considering that France was before well into the Modern Age always much more populated than all Germania (Germany plus nearby small countries). As I said previously there are serious issues regarding sampling: most of the data for Europe comes from Germany (mostly the Saale region), Sweden and a region around the Southern Basque Country (incl. Magdalenian Cantabria and Neolithic Atapuerca samples) and Catalonia-Languedoc. Recently there's been some improvement in Central Eastern Europe (Hungary, Ukraine, Russia) and there's the Chandler paper on Portugal, as well as a few isolated bits from SE Europe and France and nothing at all from Britain/Ireland (except two Paleolithic samples from Cheshire).

When I did my review on aDNA in Summer 2013, the Portuguese, Basque and German samples together added up to around 75% of the total available for all Europe. This can be improved and a lot.

Maju said...

@Krefter: ENF/EEF can't be so similar to Bedouins, if anything that could be a proxy for their West Asian ancestry but EEFs were basically identical to modern Sardinians (or if we judge by the Iberian samples, to Tuscans), being a mixture of something similar to Palestinians and something similar to WHG (UHG in Lazaridis).

Chad said...

Yes, there is some sampling issues. I feel confident that WHG rise is by local sources as it seems that WHG rise was happening simultaneously in Germany and Spain. This speaks volumes if the paper shows that.

The increase of y-I and m-U towards Central Europe is another point to make. Not only that, we have a rather steep decline WHG and ANE as you get closer to and into France. You see a 5% dip in ANE from Central Germany to East-Central France. If not for Germanic migrations, this could be a 6%+gap. It is only Northern France, with more Frankish/Norman ancestry that scores closer to Germany, yet still comes up short by about 2-3%. This all together, should point to local comebacks, with heavy mixing going on in Germany, prior to migration of R1b to the west. These horse eating Beaker folk, with the full-package, look like totally new people.

I know you don't agree with that, but snp's are pointing to much younger dates for R1b. I have someone checking snp's of MA-1, for calibrating rates a little better. This new R1b paper is probably focused on snp's, which is why they have a 10kya date.

Anyway, it should all be very clear quite soon.

Davidski said...

rk,

Making a composite out of MA-1 and a Lezgin was easy. On the other hand, extracting pure ANE segments from present-day individuals is impossible.

Masking East Asian segments in Amerindians and triggering an ANE component with them is about as close as I'll ever come to extracting ANE from modern samples.

Maju,

ENF represents the Near Eastern part of EEF. I refer you to the fateful triangle...

http://eurogenes.blogspot.com.au/2014/12/the-fateful-triangle.html

Chad said...

Maju,
The ENF is a hypothetical ancient Near East farmer, not a European farmer (EEF). This is the hypothetical farmer, prior to mixing in the Balkans and up the Danube.

Maju said...

Indeed, I just noticed when toying around with the variant plot that David posted earlier. I wonder if the Gok2 plotting is correct, because in previously published plots Gokhem samples tend to fall on SW Europeans, so I'm guessing there's some correction involved, maybe against the known problem of projecting samples on modern PCAs.

Anyways, THESE are my annotations on the Euro-PCA with ancient samples (excepted ANE apparently, which is just too remote anyhow to matter much) and something I notice is that, while the "ENF" dot is indeed realistic for the West Asian ancestors of EEFs, when we ponder the origins of Yamna people, which are said to be 50% EEF, the second population must be at the other extreme of the modern West Asian range (but with even further drift because of the ANE).

That probably implies that the bulk of the modern West Asian range already was there in Neolithic times and that the founder proto-EEF ("ENF") population should not be assumed to represent all the range of ancient West Asians but most likely an extreme. Of course the alternative explanation is that a lot of steppe genetic flow (very rich in ANE) altered the West Asian genetic landscape as it happened in Europe but most West Asians have quite low ANE affinity, so most unlikely.

I wonder how did Krefter got ANE plotted as he did. It doesn't fit well with the original data on Mal'ta affinity worldwide, which should be much stronger in Europe (at least the North) than in West Asia.

Davidski said...

Here's the PCA from the Kostenki14 paper. It doesn't suffer from projection bias as much as the Skoglund plot.

http://imageshack.com/a/img538/2013/ehUOcC.png

Gok2 plots well north of Sardinians.

Davidski said...

Maju, why did you model Yamnaya as 50% Motala12?

Yamnaya is 50% EHG, not Scandinavian hunter-gatherer (SHG).

Maju said...

Because I thought that dot was EHG and not Motala (there's no legend in that plot, although I should have noticed from the one posted by Krefter). My bad. :(

Krefter said...

Maju,
"That probably implies that the bulk of the modern West Asian range already was there in Neolithic times and that the founder proto-EEF ("ENF") population should not be assumed to represent all the range of ancient West Asians but most likely an extreme. "

People have been making the wrong assumption since LAz 2013, that west Asia and Europe got it from the same Bronze age steppe wave. It's good to see confirmation of speculations some of us have been having.

ANE in south-cenrtal Asia and west Asia have their own unique stories. I think over the next few years we'll get an idea with or without ancient genomes where it came from.

Reich, Laz, and Patterson are probably setting their sights on south-central Asia, Caucasus, and the first farmers from the near east.

Maju said...

"Reich, Laz, and Patterson are probably setting their sights on south-central Asia, Caucasus, and the first farmers from the near east".

That sounds very promising. It's not what I'm asking most (more "Atlantic" research) but also has its own interest (and less Eurocentric than my own priorities, admittedly).

I've been tempted to toy on the PCA like you but the reality is that PCAs have their limitations, as the various graphs I collected HERE from the supplements of Lazaridis et al. show.

Another issue is that while the tendency of the "Afrasian" ancestors of Stuttgart (EEF) does point towards Palestine/Arabia, if we accept the rough estimate that Laz. et al. make of Stuttgart being more "drifted" (i.e. "pulled", not actual genetic drift but rather inference of admixture) to WHG than to West Asians, the resulting ENF ancestor should be quite farther down in the plot than what you draw, probably being not similar to any modern population but as "off the range" or almost as WHG are relative to modern Europeans. That makes sense even in the absence of major demic rearrangements in West Asia (which probably happened anyhow, at least to some extent) because populations naturally evolve and also absorb genetics from their neighbors ("we are not our ancestors"). Food for thought.

Unknown said...

Maju, that thing off the plot would be Basal Eurasian. No one is full basal Eurasian. ENF is likely just over 60% Basal, whereas Stuttgart is 44%. That is the difference.

Grey said...

@Davidski

"ANE isn't a south Caspian component ...
Here's a spatial map of ANE..."

That distribution looks like it could once have been a massive circle in north Eurasia with a big corridor cut through it by a later expansion.

.

@Kristina

"In my opinion, you tend to attach far too much importance on y lines as factors of genetic divergence and convergence."

I think dramatic changes will mostly revolve around y haplogroups with mtdna working more over time.

Maju said...

@Chad: No, "Basal Eurasian" in Lazaridis et al. is something that is actually different from properly Eurasian and only found in EEF and possibly in Palestinians (where it appears as actual African admixture).

There is the unsolved issue of this BEA thing is actually a residue of the populations who left Africa for Arabia before migrating to South and SE Asia (where they became "true Eurasians" by founder effects and remixture) or it is actually African admixture from, probably, the Nile region (what I consider more likely, particularly because of the importance of Y-DNA E1b in the process and also because in some algorithm-generated trees, Skoglund supp. materials, it seems to be Dinka-like). But, regardless of what exactly it is, BE is only a small fraction of the overall non-European element in EEF (ENF in Krefter's terminology) and not the same thing.

What we know is that the non-European element (non-UHG or, by proxy, non-WHG) in EEF tends towards Palestine/Arabia but we don't know where the "source dot" actually is.

But never mind me because I was wrong, I realize now: what the phrase from Laz. et al. implies is just that the PCA plotting does not reflect the real distances, because WHG are closer to EEF (and therefore also to modern Europeans) than these are to West Asians. We still don't know where the "source dot" ENF would be plotted if we knew it but what we know is that the distance along the WHG-Palestine line is wrongly expressed in the PCA, because it appears that EEF are closer to Palestine than to WHG, when it's actually the other way around.

This seems to be a serious problem inherent to PCA plotting. Other means are needed such as f4 tests to really measure these admixtures in realistic scores. So all the exercises we do on PCAs will unavoidably be wrong to some extent. PCAs after all only measure things along two axes and those axes are defined by the main polarities among the majority of samples involved in their creation (not the ones projected afterwards, nor those with minority peculiarities either).

Davidski said...

Maju, early European farmers are closer to modern Near Easterners than to Western European Hunter-Gatherers (WHG). That's because they're at least 60% Neolithic Near Eastern, and probably around 70%. But they are closer to WHG than to Ancient North Eurasians (ANE), because they're mixed. On the other hand, WHG are closer to ANE.

And yes, it's possible to read admixture proportions from that PCA, because that PCA is based on admixture proportions and not genotype data (from which it's a lot more difficult to read admixture proportions).

But it's obviously easier just to consult the K8 spreadsheet.

Matt said...

@ Davidski Here's the PCA from the Kostenki14 paper. It doesn't suffer from projection bias as much as the Skoglund plot.
http://imageshack.com/a/img538/2013/ehUOcC.png
Gok2 plots well north of Sardinians.


David and all -

Funnily enough, it sort of doesn't suffer from projection bias at all, because actually none of the samples are exactly projected, where by projected we mean "the PCA was run without any ancient samples, then the ancients were projected on the PCA".

What Seguin-Orlando et al did instead was run a PCA with one of each of the samples and then procrustes fit all the PCA together.

p 26 For each ancient genome, Principal Component Analysis (PCA) was performed on the matrix of haploid genotypes for all SNPs with data in the ancient individual, using the R function svd. These individual PCAs were then combined using procrustes transformation, as implemented in the R package vegan (http://cran.rproject.org/web/packages/vegan/index.html) .

p 66 Fig. S23 - Procrustes PCA of all ancient genomes on modern populations from Europe, Middle East and Caucasus. Bold letters indicate population average PC.

Effectively IRC this means each non-ancient sample shared between all the PCA gets a consensus position based on all the PCA together, then the ancient samples are positioned relative to them based on their position relative to the shared samples on their own PCA.

In a sense I think this is very arguably the most accurate result, as it doesn't allow any drift like errors or individual subpopulation variation in the individual ancient samples to hijack the plots.

This sort of non-projected procrustean PCA would be very cool to see using the Eurogenes PCA with all the ancient samples together.

As far as I can tell you just need the output of the original PCAs to go through the Procrustes function and you can place all the samples together on a plot. Then we can have one nice big plot for Eurogenes readers showing the truest genotype based location of all the ancient samples together, including samples not plotted together by Seguin-Orlando et al, such as Oetzi and KO1.

Maju said...

You're right David: in table S10.2 the Bedouin-like admixture apportions in Stuttgart are estimated to be between 62-98%, depending on how bigger is the estimated fraction of African ancestry in Bedouins - the larger one, the larger the other. For African admix. in Bedouins between 4.2% and 7.2% (but what if <4.2%, i.e. if Bedouins actually represent the "real thing" or something very close?) It seems I read something wrong in any case.

Krefter said...

"That sounds very promising. It's not what I'm asking most (more "Atlantic" research) but also has its own interest (and less Eurocentric than my own priorities, admittedly). "

100s of Iberian genomes and mtDNA samples from the Mesolithic-bronze age will be published in the next few years. If you look through recent Eurogenes posts, you'll see the abstracts. Also, Reich got Neolithic genomes from Spain.

I don't know of anything from the British isles or France, except the "Amesbury Archer" who was a Bell Beaker man from England.

Davidski said...

Check out the WHG scores for these Tajiks.

Pop ID ANE South_Eurasian Near_Eastern East_Eurasian WHG Oceanian Pygmy Sub-Saharan
Tajik_Pomiri tdj126_shugnan 0.30497 0.097263 0.422331 0.024766 0.127042 0.011101 0.00001 0.012517
Tajik_Pomiri tdj252_ishkashim 0.31573 0.097505 0.409801 0.044426 0.092124 0.017065 0.004825 0.018525
Tajik_Pomiri tdj255_ishkashim 0.315185 0.127368 0.445983 0.00543 0.071769 0.018749 0.006968 0.008548
Tajik_Pomiri tdj269_ishkashim 0.317442 0.09198 0.416025 0.043549 0.101742 0.018908 0.001155 0.009199
Tajik_Pomiri tdj305_shugnan 0.30346 0.096717 0.415781 0.042278 0.132453 0.0079 0.00001 0.001402
Tajik_Pomiri tdj508_rushan 0.285153 0.086952 0.447696 0.052695 0.115301 0.003692 0.00001 0.008502
Tajik_Pomiri tdj868_rushan 0.280677 0.082906 0.432444 0.037771 0.141384 0.013681 0.002565 0.008572
Tajik_Pomiri tdj979_shugnan 0.303188 0.086313 0.450931 0.027539 0.102511 0.011411 0.000157 0.01795

Chad said...

Yep, that's what I was telling you about about 3-4 weeks ago. You see the same in I believe Kyrgyz, and Uzbeks. You can see the WHG from Andronovo, fairly clear. That's why I as saying that they look about 1/4-1/3 Yamnaya, and maybe up to 1/2 max of Andronovo.

Chad said...

It might have been Turkmen too. Kyrgyz may have been on the lower end. I can't find that k8 spreadsheet anymore. Could I please get the link, David? Thanks!

Chad said...

Take East Eurasian out of the more Turkic groups, and you see a similar ratio of the remainder that belongs to WHG.

Maju said...

"I don't know of anything from the British isles or France, except the "Amesbury Archer" who was a Bell Beaker man from England".

Born in Switzerland probably. But there are two ancient mtDNA samples from Cheshire since long ago: one was U5 and the other R-CRS (possibly H but uncertain), they are listed at Jean Manco's site (with all her typical caveats). However nothing from recent times. The few French samples have already been commented.

"100s of Iberian genomes and mtDNA samples from the Mesolithic-bronze age will be published in the next few years".

Sounds good.

Krefter said...

People knowledgeable in archaeology,

The small amount WHG in some Indo European-speaking Asians is still very confusing to me.

Considering Yamna's Near eastern ancestors may have been Armenian-like, and some might not have been from the Balkans, is it possible there was a very near eastern-like population outside of Samara(Like chad has suggested), which had many descendants who spread IE languages in Asia?

I haven't investigated whether ENF is sucking in WHG yet, but that may be apart of the problem.

Maybe IEs were simply much more successful in terms of genes in north Europe than else where.

Andronovo looks very EHG and Yamna-like to me, so that could be the case with modern Indo Iranians in Asia.

I mean look like Turkish people, they have very weak signals of north Asian ancestry but speak a Turkic language.

Chad said...

This ENF pretty much solves that WHG problem.

It is still easy to pick out the WHG relation to others. If someone could provide me the link to the full k8 spreadsheet, I'd be happy to go over it a bit.

Krefter said...

Chad,

https://docs.google.com/spreadsheets/d/1kd9Q9vFrL1Cra9ayqMYVFKXrUdnThmQJVMtjczLhoTs/edit#gid=2065345772

Seinundzeit said...

The Pamiri Tajik results are exceedingly interesting. In the three-pop results that were done awhile ago, they scored Pashtun+Loschbour, Burusho+Loschbour, and Kalash+Loschbour. But their ANE signal was comparable to other South Central Asians. All of this is evident in their K8 results, pretty much identical to other South Central Asians, but very WHG-shifted.

Nirjhar007 said...

@Krefter
''Considering Yamna's Near eastern ancestors may have been Armenian-like, and some might not have been from the Balkans, is it possible there was a very near eastern-like population outside of Samara(Like chad has suggested), which had many descendants who spread IE languages in Asia''
Anthropology also agrees that folks of Armenian area bear strong relation with Yamnaya,CWC,Andronovo folks! as i linked in the previous discussion.

Maju said...

@Krefter: in the Kurgan model, the Indo-Iranian branch derives from Andronovo. Ultimately this culture has its roots in Poltavka, which is heir of Eastern Yamna. Not sure if the Yamna samples are Eastern, from the Volga, or Western, from the Don-Dnieper area; they may have got some differences and, if they are Western, they may not truly reflect the inheritance of the Eastern population, which is the one that granted continuity in the long run. These kind of hypothetical differences would not affect Europe (as one would expect Western Yamna to be more closely related with Balcanic and Central European kurgan groups) but MIGHT affect the legacy in the steppes and hence what affected Southern Asia later on.

Beyond that origins issue, there's another issue, which is if the Kurgan invasions in India and such had really a major impact in terms demographic or were rather just a thin aristocratic layer, which mostly fed from the pre-existent IVC and other Neolithic cultures. The main first IE culture in South Asia is not really an independent one but rather the last layer of IVC: cemetery H, in which customs are altered (horse sacrifice, cremation) and unprecedented violence is apparent. This can be interpreted as conquest by small groups of nomadic warriors rather than replacement.

If there was no replacement (almost), you'd see no genetic signal (almost).

But anyhow, something apparent in Central Asia (and also in the area of Samara, what seems important) is that Y-DNA R1a is not of the European branch but of a distinct Asian one, Z93. Unlike in Europe R1a in Asia does not look too related to IE expansion (at least not within the Kurgan model) but, in any case, it is so strikingly different that it suggests that the populations involved in Andronovo were not really descendants from the Western-Yamna-like populations but something only vaguely related, maybe peoples from a different origin assimilated to IE culture.

Overall (and I have all kind of doubts, as you can infer from my words), it's very possible that the Indo-Iranian expansion only has a cultural and not much genetic relation to that of Western IE (or even, as we have recently seen Tocharians/Afanasevo).

Anyhow, I tend to think that early IE migrations had much greater impact, particularly in low density areas like those around the Baltic, because they still partly belong to the Neolithic (Chalcolithic) and not fully to the Metal Ages (Bronze & Iron). As we get into the real Metal Ages, the we see more conquest and less mass-migration, an aristocratization of these violent processes in which the important goal was to conquer not lands to work, but lands with their native workers intact and alive (albeit subjugated). We can see this trend still in the Middle Ages and I think that the Altaic migrations offer a good example of how complex could be the older but very comparable IE ones and how conqueror populations tend to nearly disappear in densely populated conquered areas, even if they can assimilate them, for example Turkey.

Krefter said...

The Yamna sample are from Samara, Russia, which is at the eastern end of Yamna.

Nirjhar007 said...

@Maju
There is no Kurgan Migration to India the Cemetery H bears the catastrophe of Climatic events and Social Distress effected by that nothing else, horses were also found from other Indus Sites and Cremation is irrelevant as In Rikveda and Shatapatha Brahmana burial is comprehensively discussed which are of Mature harappan phase etc!
So, Just Shut up!!!!

Nirjhar007 said...

Check out even in the case of Afanasavo Culture IN South Siberian Area it has been pushed back to ~4000 BC!!!
http://en.wikipedia.org/wiki/Afanasevo_culture

Krefter said...

Thanks for the info Maju. I agree with Davidski and Chad that there are signals of Yamna-like ancestry in central Asia, and that it probably comes from Indo Europeans. I doubt Andronovo was very different from Samara Yamna, but who knows.

The aristocrat idea makes sense. The huns left little to no descendants in Europe, even though they ruled most of it for 100 years or so. The Turks in Anatolia is another example.

Nirjhar007 said...

@Krefter
Seeing that comprehensive and cool Chart I think Armenian-Iranian-Pathan population can be modeled as the pop which entered Yamnaya ~3000 BC?

Chad said...

It looks like Kazakhs were definitely over 20 and maybe up to 30% WHG, prior to East Eurasian arrival.

Kyrgyz are 6% WHG with much more East Eurasian than Tajiks. They could've been close to 20% WHG, before Turks came in.

Uzbeks could've been close to 20%, as well. It's a pretty similar pattern.

Chad said...

I think Andronovo will be 40%ENF/ 25%WHG/ 30%ANE/ 5%EEUR

Nirjhar007 said...

@Chad
''It looks like Kazakhs were definitely over 20 and maybe up to 30% WHG, prior to East Eurasian arrival.

Kyrgyz are 6% WHG with much more East Eurasian than Tajiks. They could've been close to 20% WHG, before Turks came in.

Uzbeks could've been close to 20%, as well. It's a pretty similar pattern.''
What about their high East Eurasian Ancestry?

Nirjhar007 said...

@Chad
''I think Andronovo will be 40%ENF/ 25%WHG/ 30%ANE/ 5%EEUR''
Do we have K8 on Scythian stock also Chad?

Chad said...

I think their East Asian ancestors were like the Yakut. The previous paper on Turks supports that. Same region.

Nirjhar007 said...

Chad, Andronovo is not Aryan its Pre-Scythian its certain!

Chad said...

David,
I'm doubt that this can fit on the triangle, but if you would..

15%ANE
25%WHG
22%ENF
9% SEUR
29%EEUR

This is probably what my daughter would score. Thanks!

Chad said...

Nirjhar,
I see no point in arguing it. Someone brought WHG towards South Central Asia, and they had a good amount of it. They weren't from Iran.

Davidski said...

Here you go...

https://drive.google.com/file/d/0B8XSV9HEoqpFY0dXTm5kWkp6Um8/view?usp=sharing

https://drive.google.com/file/d/0B8XSV9HEoqpFRkZEbkJRc1BrRkU/view?usp=sharing

Unknown said...

Awesome! Thanks!!

Krefter said...

Davidski

Wait, so east Eurasian makes an effect on the ANE K8 PCA? If so, can you show me where pure east Eurasian places? When I projected myself on a Microsoft Paint ANE K8 PCA I was west of where you had me, is it because of east Eurasian and Sub Saharan?

Davidski said...

Yes, East Eurasian and Sub-Saharan does affect your position in the triangle. East Eurasian will pull you to the east. Have a look at the biplot.

https://drive.google.com/file/d/0B9o3EYTdM8lQSXFBLTU1QXRHVGM/view?usp=sharing

But there's no point plotting it on a West Eurasian plot.

Nirjhar007 said...

@Chad
''Nirjhar,
I see no point in arguing it. Someone brought WHG towards South Central Asia, and they had a good amount of it. They weren't from Iran.''
Yes but that is not the issue i'm talking i'm talking about the identity of Andronovans as depicted in Aryan Scriptures as people of Nomadic lifestyle, fond of horses they also had a name as 'Tuiriyas' which remained as Turanian in Iranian dialects....
About WHG they acquired it from N-W Eurasia but i suspect majorly by Mtdna not Y-DNA!

Unknown said...

Yeah, i was thinking she might get compressed a bit since she's about half Asian and Native American.

jackson_montgomery_devoni said...

A while back through email David Reich told me that all of the Mesolithic Samara hunter-gatherer mtDNA samples that they tested belong to haplogroup U except for one that belongs to H. Subclades are what is important here. So far no H1 or H3 have shown up in pre-Neolithic European remains.

Maju said...

If you check them carefully, one of the samples of Chandler, Sykes & Zilhao 2005 for Epipaleolithic Portugal must be H1b. So what you say, Jackson, seems incorrect.

Anyhow, most ancient mtDNA sequences included only HVS-I, which fails particularly at identifying H of H1, so it's only normal that these lineages have not been properly identified yet (even if a large array of R*-CRS and other R* could well be precisely that). Even recently some studies like Hervella's only complemented this methodology with RFPLS, which is enough to detect H (what she did) but not to detect H1.

So your claim, Jackson, is nothing but a fallacy.

Chad said...

I think the issue with that H1b was an overlap of archaeological layers with burials. I'll check on that later.

Unknown said...

http://dienekes.blogspot.com/2015/01/55-thousand-modern-human-from-manot.html?m=1

Could be pre-Neanderthal admixture. Let's hope someone can grab some aDNA out of this one.

Marnie said...

@Chad,

"Could be pre-Neanderthal admixture."

What are you drinking? The "pre-Neantherthals" would have to be from more than 300,000 years ago.

This one is from 40,000 years ago.

Maybe it's time for you to get off the blogs and TED talk circuit, and read the last five years of literature on archaic admixture.

Unknown said...

Marnie, I'm talking about the dated mixture between UP and Neanderthal. That should be obvious. Many feel confident in the range shown by Ust-Ishim.

Marnie said...

This sample is quite a bit after the estimated range of dates for Neanderthal admixture.

It's quite far south from Ust-Ishim.

Both MA1 and K14 have more detectable archaic admixture than humans today.

So far, we have only a few Neanderthal and Denisovan samples that we can use to check for admixture.

I doubt that this sample in Israel would have no archaic admixture. But it would probably be different archaic admixture than MA1 and K14.

Chad said...

The admixture is expected to have taken place in the Near East, at about this time. So, it will be perfect. You need to go back and look at the Ust-Ishim paper.

Chad said...

Here....
This UP Near Easterner would be perfect to find the mixture date.
From the paper:

However, the genomic segments of Neanderthal ancestry are substantially longer than those observed in present-day individuals, indicating that Neanderthal gene flow into the ancestors of this individual occurred 7,000–13,000 years before he lived. We estimate an autosomal mutation rate of 0.4 × 10−9 to 0.6 × 10−9 per site per year, a Y chromosomal mutation rate of 0.7 × 10−9 to 0.9 × 10−9 per site per year based on the additional substitutions that have occurred in present-day non-Africans compared to this genome, and a mitochondrial mutation rate of 1.8 × 10−8 to 3.2 × 10−8 per site per year based on the age of the bone.

Unknown said...

Oops, this would be pre-UP, right?

Marnie said...

Various papers estimate dates for Neanderthal admixture as between about 70,000 and 54,000K years ago (approximately).

I doubt that we have the full picture of Neanderthal, Denisovan and other types of admixture, as yet. And I also don't think that archaic admixture was a single event in time or place.

We also don't the range over which Neanderthals, Denisovans and other archaic hominins overlapped in time and place with modern humans.

It just isn't helpful to assume that Neanderthal admixture took place (only) in the Near East, only in one place, or only in a narrow time window.

We don't know.

In any case, it's unlikly that the DNA from this cranium will be able to be sequenced.

Marnie said...

Correction to previous post:

54,000 years ago (not 54,000K years)

Alberto said...

@Marnie

"It just isn't helpful to assume that Neanderthal admixture took place (only) in the Near East, only in one place, or only in a narrow time window."

I once read an article (unfortunately I can't find it now) that explained with good logic that if this admixture event took place, indeed it has to be a single event. It argued that all non-african modern humans are descendant of a female who was born to a Neanderthal father and a Sapiens mother. The offspring of this single female seem to have had some big competitive advantage, not only over Neanderthals, who got extinct, but also over "pure Sapiens", who also got extinct among the non-African population.

If I find the article I'll post it so people can check the arguments (which seemed quite sound to me at the time).

Alberto said...

Here:

https://constantincretan.wordpress.com/2011/09/12/the-first-cro-magnon-was-a-hybride-eve-with-a-homo-sapiens-mother-and-a-homo-neanderthalensis-father/

Marnie said...

@Alberto

"I once read an article (unfortunately I can't find it now) that explained with good logic that if this admixture event took place, indeed it has to be a single event."

I doubt that. Anyway, again, we don't know for certain.

Here's a link to Svante Paabo's recent talk on ancient admixture:

http://theolduvaigorge.tumblr.com/post/103116348132/a-neandertal-perspective-on-human-origins-part

Alberto said...

@Marnie

"I doubt that."

Why? What are your refutations of the points explained in the article?

Of course we still don't know for sure (we need factual data for that), but applying logic to the data available is better than just wild guessing without any logical argument.

The talk you linked to agrees with the basic data points used in those arguments (at least for the part I had time to watch). Better link here: https://www.youtube.com/watch?v=8flcCtIkTUc

Marnie said...

@Alberto

In the talk that Svante Paabo gave at Oxford in the Fall, he mentions that admixture from Neanderthals was probably from both male and female Neanderthals.

The talk is in three parts. It's about an hour long.

Anonymous said...

@ Maju
r1a among indo-iranians is not very diverse and among dravidians it is even less diverse so the spread of r1a-z93 must be the result of a founder effect among indo-iranians and can not be associated with dravidians who have much less of it than indo-iranians. Indo-iranians have much r1a-z93 from early eastern indoeuropeans but not much autosomal dna from them because the patriarchal indo-iranian tribes took all kind of foreign women during their migrations but tried to preserve their patrilineal lineage. R1a-z93 is too young and was already present among people of the andronovo culture who must be very close to the yamnaya people but with more native central asian admixture. R1a-z93 is actually also present in europe in very small numbers but the european subclades of r1a are almost absent in asia .This must be the result of a founder effect among indo-iranians and r1a-z93 is just a sister subclade of balto-slavic r1a so there must be a ethnic and also linguistic connection between carriers of r1a-z93 and european r1a

Marnie said...

@Alberto

In the question/answer period at the end of Svante's presentation, he discusses the issue of male/female Neanderthal admixture.

It's in part 3 at about 5 minutes into the video.

Maju said...

@Skilur: within Z93 and Z95 the SNP-structured basal diversity in India is high, however not enough to consider it the most likely origin of Z93, which can well be a bit to the NW somewhere around Afghanistan (but not necessarily within Afghanistan). Altai and Central Asian lineages (including the Samara area) are derived and can't be at the origin of South Asian R1a. No apparent N→S flow is detectable in the STR structure (once sorted by SNP haplogroup hierarchy).

I discussed all this (and some other related stuff such as the ages of Y-DNA R1) here: http://forwhattheywereweare.blogspot.com/2014/03/y-dna-r1a-spread-from-iran.html

... but you may just want to look at this map, which is my synthesis of R1a spread on Underhill's data.

Pushing the estimate dates forward a lot, one could make the "Afghan" expansion of R1a-Z93 be Copper Age, but then R1a could not be in Europe in Yamna times or, as we know for a fact it was, in Afanasevo either (both are c. 2000 years older and require a previous development of some sort, so add a couple millennia), so this is a practical impossibility unless Underhill 2014 is seriously challenged (something I have failed to read in a comprehensive and convincing format of any sort - just short low-coherence protests, often one-liners, that don't seem enough at all).

Hence it seems much more logical to think that Z93 expanded in Asia in the Neolithic or Mesolithic and that the European branch also did (from a different origin surely), regardless of some important later re-expansion with Kurgans, which seems limited to Z282, and hence to some specific parts of Europe (Poland-Russia).

Alberto said...

@Marnie

Thanks. He mostly agrees that it's probably mostly from male, but he shows some doubts. Of course, it's difficult to say with the data available.

I still think that the reasoning in the article I posted is quite good to support a single event and from a male Neanderthal and female Sapiens. At least it seems the most likely, because anything else would require very unlikely combinations to happen.

Unknown said...

A little more background..

http://www.theguardian.com/science/2015/jan/28/ancient-skull-found-israel-sheds-light-human-migration-sex-neanderthals

Seinundzeit said...

David,

If it isn't too much trouble, could you run these three Kalash samples?

HGDP00315

HGDP00286

HGDP00290

Thanks in advance.

Marnie said...

@Alberto

"I still think that the reasoning in the article I posted is quite good to support a single event and from a male Neanderthal and female Sapiens. At least it seems the most likely, because anything else would require very unlikely combinations to happen."

Paabo specifically says that Neanderthal admixture is *not* from a single baby.

It's quite subtle and based on your comments, I think you're missing the subtlety.

Krefter said...

Davidski,

rms2 at Anthrogenica is starting a red hair-variants project at FTDNA sometime in the future.

Can you test the Hinxton Celts and Anglo-Saxons, Otzi, all the Goks and Ajvs, and all the motalas, for these red hair variants?

http://www.anthrogenica.com/attachment.php?attachmentid=3617&d=1422406771

Davidski said...

Just list the samples and SNPs here, and I'll put their results in a spreadsheet.

Krefter said...

Here are all the samples and SNPs in a spreadsheet. One set is of samples who don't have results on just one SNP(rs34474212) and the other set is of samples who don't have results in any of the SNPs.

https://docs.google.com/spreadsheets/d/1MjlTxOTAb7FX89lWm4F9A9RGFPtvekbnGljfnzt4neU/edit#gid=0

Thanks.

Unknown said...

@Davidski --

The high points in your ANE map look like a combination of North Caucasian, Burushaski, and Yeniseian, which is the hypothetical macro-family "Sino-Caucasian" minus the Sino part.

Grey said...

@Alberto

"a male Neanderthal and female Sapiens. At least it seems the most likely, because anything else would require very unlikely combinations to happen."

I'd have thought a single event over thousands of years would be extremely unlikely.

Seinundzeit said...

Colin,

I've been arguing for this association, ever since I realized that the most ANE-admixed Eurasians were the Burusho (which was evident with David's first supervised ADMIXTURE experiment involving MA1 and Eurasia, and with Everest's MA1+Eurasian PCA plots).

Now that we know that the early Indo-Europeans were just too genetically European to have contributed any substantial ancestry to West and South Asians, we still have to explain why South Asians consistently have some of the highest levels of ANE admixture in Eurasia. The fact that, after South Asia, ANE peaks in the northern Caucasus and central Siberia, must point to the older macro-family connection. It's the only plausible connection tying all three regions together.

Anonymous said...

@ Maju
just ask Davidski he also see r1a-z93 as the result of indo-iranian expansion from the steppe. The samples of ancient steppe people had in some cases almost entirely r1a-z93 and so much higher percentages than any modern central asian and south asian indo-iranian population, who mixed much with pre-indo-iranians there.
R1a-z93 is simply too young and to similar to european subcaldes so it is impossible that it was carried by dravidians or other non-indoeuropeans who had no connection with indoeuropeans genetically and linguistically. If another subclade of r1a was carried by early indo-iranians we would find at least some traces of this subclades among very patriarchal people like brahmins but they have only r1a-z93 which is much less diverse among them compared to europeans. If r1a was of pre-indo-iranian origin it would be much more diverse and other subclades would exist but r1a-z93 there is not diverse and looks like the result of a massive founder effect.

Alberto said...

@Marnie

Yes, he says the admixture probably is from one population, but not from one single baby. He admits that most likely (but not surely) is from male Neanderthals only (or mostly). But in any case he says these things with the caution of not being able to know for sure. Just as his best guess.

I'd rather YOU read the article I linked to, and tell me YOUR arguments for doubting that it is a single event, to explain them based on logical/mathematical models, or at least to point the weaknesses in the arguments in that article. That would be much more helpful.

@Grey

"I'd have thought a single event over thousands of years would be extremely unlikely."

Yes, at first I'd have thought the same. But one day I read that article I linked above and now I think that any other possibility than a single event is actually much more unlikely. If you would read it and still think otherwise, I'd like to hear your counter arguments (with your possible model).

Nirjhar007 said...

@Skilur
''just ask Davidski he also see r1a-z93 as the result of indo-iranian expansion from the steppe.''
Yes and pathetically wrong.....

Grey said...

@Alberto

Fair enough. I could believe human-neanderthal mixture among a population just outside Africa could lead to a new breed. In fact it seems quite likely.

My argument was with the idea that there was only one archaic admixture event as opposed to one specifically neanderthal admixture cos Denisova for a start (and imo that big red ANE signal also.)

postneo said...

@chad@nirjhar

I would think the time depth of WHG makes it hard to pin point a specific geographic origin. How do we know if WHG simply survived better in northern europe but otherwise had a much wider spread? At one point R1b was thought to be west european.

@skilur
http://yfull.com/tree/R-Z93/
Check this out. z93 is an old and large branch, You cant say a population is bottle necked because they are z93. South asian samples are labeled ITU, BEB, STU, GIH, PJL. They are not very representative but at least geographically disparate. They appear all over the tree with no particular structure.

My impression: z93 from central asia is from the historic period but older central asian samples are not.

Please provide some references.

Over Central asia is a recipient and transfer zone for european or asian lineages but not a source.

capra internetensis said...

@Marnie:

thanks for the link to that blog, it is cool.

@Alberto:

The argument you linked to does not seem convincing to me.

Last year's studies found that there is much less Neanderthal ancestry on the X chromosome than the autosome as a whole, suggesting that Neanderthal X chromosomes might result in hybrid sterility. In that case we would expect hybrid daughters to pass on most of the genes, rather than hybrid sons, so Neanderthal Y DNA would be rare or absent, though mtDNA wouldn't.

However, uniparental haplogroups can easily be lost entirely due to drift in small populations. Selection would do the job more certainly, of course; if the Neanderthal mtDNA was slightly deleterious we would expect it to be lost over time while beneficial autosomal genes were retained. There is no need for a scenario where Neanderthal uniparental DNA was *never* present even for one generation.

(And the author's theory about MH having more vocal ability while Neanderthals contributed to intelligence based on modern populations seems quite ignorant of both linguistics and cognitive research.)

Nirjhar007 said...

@Postneo
''I would think the time depth of WHG makes it hard to pin point a specific geographic origin. How do we know if WHG simply survived better in northern europe but otherwise had a much wider spread? At one point R1b was thought to be west european.''
The main problem is lack of aDNA from Asia and the failure to establish a reliable method to age the admixture of a specific autosomal components in modern populations that carries it...

Michael said...

@Alberto

The arguments in the article you link to are easily disproven by the most basic genetics.

The only reasoning is that modern humans have 100% modern human mtDNA and Y-DNA. Based on this, they conclude that a single event between a male Neanderthal and a female modern human (yielding only female offspring) was the only possibility.

However, let's suppose that the only event was between a male modern human and a female Neanderthal, and the offspring was one male. This child had Neanderthal mtDNA and modern human Y-DNA.

If he had only daughters with a modern human, and those children integrated into the modern human society, then his daughters would have only modern human mtDNA, and all of their sons would have modern human Y-DNA.

Given the fact that only a very few Y chromosome lineages from that time actually survive to this day (<10). It is entirely plausible that there were several events, and the sex specific markers were quickly lost through drift.

Maju said...

@Skilur: the problem of languages' origins beats me to some extent and seems relatively irrelevant because it's extremely hard to reconstruct even the earliest Holocene relationships. It's perfectly possible that the same general lineage had founder effects associated to two different linguistic families anyhow. Not something I'd base my judgment on, really.

So back to the genetic aspect, we have just read how Afanasevo burials are R1a(xZ93), in spite of this other branch being today effectively absent in the region. This is just one data point but it clearly indicates that Z93 was not implicated in the early IE expansion in Central Asia in an overwhelmingly dominant form.

But more importantly Z93 is almost the age of R1a as such. There's no significant time difference between one expansion and the other. Using Méndez' Anzick-calibrated mutation rate, the lineage spread c. 7300 years ago, what is coincident with the Neolithic spread from West Asia to its peripheries: Europe, South Asia and Central Asia. I personally have another recalibrated estimate that would produce an Epipaleolithic date of c. 11,000 BP for R1a and c. 10,000 BP for Z93 - I may be wrong however but the Méndez-rate estimate should be very persuasive.

Of course you may adopt a "conservative" stand, as Underhill did, and just look for the average of scholastic mutation rates, what allows for a 6000 y.o. IE expansion... but not from Samara, only from Iran or Kurdistan. And that doesn't make the slightest archaeological sense. The estimate that fits best with a Zagros-area expansion of R1a is clearly one that puts it in the Neolithic time-frame. And that is not my own recalibration, which may go a bit too far, but the Méndez-rate estimate, which the only one calibrated to a real aDNA sequence (Anzick), and hence the most scientifically plausible.

What languages those people spoke beats me. People do not only speak the language of their forefathers but actually change languages in very little time (one or a few centuries) given the right circumstances. Languages are tools and people adopt them according to their needs.

In Europe anyhow, we'll see what the Yamna lineages actually are when published, but from modern data it seems R1a-Z282 did not expand from the Volga at all but rather from somewhere around Kiev. This means that, even if it is a founder effect associated to IE/Kurgan expansion, it is not necessarily something that Kurgan invaders brought from the Volga, but rather something they picked on their way to Germany and Poland. This is underlined by the fact that, in spite of the clear impact of early Kurgans in the Eastern Balcans, there is very little R1a of any kind over there.

Anonymous said...

@ Maju
afansevo was most likely not indo-iranian and had maybe something to do with the tocharians, who spoke indo-european languages least related to indo-iranian languages so it makes sense that they carried another subclade of r1a . Also r1a is not so old like you wrote and iran and kurdistan have quite low percentages of it. There is no reason to asume that it orginated from modern iran or kurdistan. R1a-z93 is just too dominant among the most endogamous indo-iranian population and turks in northern central asia and ancient indo-iranian people there carried mucg r1a-z93 in regions where never any south asian or west asian culture had any important impact (mongolia,siberia,...)
https://www.familytreedna.com/public/r1a/

I dont see any reason why r1a-z93 can not be the result of any founder effect and the intern diversity of r1a-z93 is the result of several waves of indo-iranian tribes which migrated to south and central asia (scythians,indo-aryans, iranians,...). But other subclades of r1a are almost non-exiting that because of a founder effect among early indo-iranians and other indo-europeans didnt migrated to southern central asia and south asia. The kushan and yuezhi were actually most likely iranic tribes unrelated to tocharians in the tarim basin so because only indo-iranians settled in afghanistan,iran and india only r1a-z93 can be found there

Alberto said...

@capra internetensis
@Michael

I don't want to keep this off-topic conversation too long here. But in that article there is not only one, but several arguments. It is the combination of them that makes the scenario the most likely (though it cannot be 100% proven, of course).

Rebutting one argument alone is not enough. There are possible scenarios where mtDNA could be lost for other reasons, or Y-DNA could be lost for some reason. But combined with other facts (like only a maximum of 50% of the Neanderthal genome being present on non-African humans, lower contribution to the X chromosome than the autosome, the retaining of 1-4% of the genome throughout thousands of generations, the non existence of "pure" Sapiens outside Africa,...) does lead to a conclusion that a single event, from a male Neanderthal and a female Sapiens is the most probable (other possibilities would require very complicated circumstances in a very complicated order so that we would end up with today's results in living humans).

Again, it's off-topic and not important. Time will tell whatever happened (if it's possible to ever know).

(P.S: I don't what you mean about vocal ability of MH. I don't know who's the author of the article, and certainly won't support any of his other theories in other articles that I have not read. But however crazy they might be, they have nothing to do with the arguments in the article we're discussing.)

Maju said...

@Skilur: I know what you say about Afanasevo but I don't agree with your conclusions, which seem illogical. Basically you're implying that, even if Tocharian is a historical language family which co-existed with Indo-Iranian in the steppe, that somehow Z93 IIs displaced pre-Z93 Tocharians. I say this makes no sense, that what the only thing clear is that R1a(xZ93) Afanasevo people had a negligible genetic impact long term in their area of ethno-linguistic influence (in fact Uyghur R1b has traditionally been assumed to be the main legacy of Tocharians, although this still requires of confirmation of some sort).

" But other subclades of r1a are almost non-exiting that because of a founder effect among early indo-iranians and other indo-europeans didnt migrated to southern central asia and south asia".

What we see is the opposite: that all subhaplogroups (and higher phylogenetic subhaplotypes in general) are present in the south (India-Iran) but not in the North (Central Asia/Siberia), that the only apparent trail of migration is from South to North. More precisely: two branches to approx. Uzbekistan and Altai each and then a derived merge of them in Samara (which one could well attribute to Turkic rather than IE expansion - but this hypothesis needs some other support which so far I don't know of).

We see even one branch that appears to migrate from India to Central Asia, although the bulk seems to rather originate towards Iran. But we don't see any lineage that appears to migrate from Central Asia (let alone Europe) to India or Iran. So again we are before a case in which the genetic impact of the IE conquerors seems negligible to the point of not being detectable.

Unless you argue for an IE expansion in the Neolithic, one from Iran-Kurdistan-Anatolia, there's no way to insert all R1a in it. Only Z282 seems to fit in and even that one seems to have hijacked rather than truly originated the Kurgan expansion, with a very specific Pitted Ware/Corded Ware type area of impact towards Central-East Europe. Probably Uyghur R1b also hijacked the Afanasevo/Tocharian expansion rather than being the core of it. And if some lineages did the same in a South Asian context, these must be lineages that were already in IVC, including but not only R1a (also L, J2, etc.)

Krefter said...

Maju, so you're saying R1a-Z282 was picked up by IEs while moving into Europe and R1a-Z93 was picked up by IEs while moving into Asia?

Isn't that a bit coincidental? How do you explain Andronovo and its descendants being practically uniformly R1a, and mostly R1a-Z93(based on STR haplotype)? Add to that Andronovo mtDNA wise obviously being very similar to Yamna, like Corded ware(where R1a is popped up in its only samples)?

If no R1a is found in Yamna, I congratulate you(it doesn't confirm Yamna didn't spread it, but is evidence), for sticking to a theory most didn't believe.

"This is underlined by the fact that, in spite of the clear impact of early Kurgans in the Eastern Balcans, there is very little R1a of any kind over there."

The Balkans probably have much less Yamna-like ancestry than northeast Euros. Also, not all eastern waves had to of brought R1a.

Davidski said...

Krefter, I only have one of those SNPs in both of my datasets.

https://drive.google.com/file/d/0B8XSV9HEoqpFd082R3puc0RKNlU/view?usp=sharing

Maju, R1a-Z93 is an Indo-Iranian marker which first expanded with the proto-Indo-Iranian Andronovo culture. It arrived in South Asia well after the Neolithic in several waves, and the reason it's so common there is because of a series of massive and rapid founder effects (which is also why its STR diversity is so high there).

Maju said...

"... so you're saying R1a-Z282 was picked up by IEs while moving into Europe and R1a-Z93 was picked up by IEs while moving into Asia?"

Not exactly. You don't seem to get my nuances nor my relative uncertainty to the detail of what actually happened.

I'm saying, for Europe, that the "geo-structure" of Z282 looks like expanding from around Kiev in a way that could be partly consistent with Corded Ware, particularly the expansion into Poland and East Germany of (mostly) its subclade M458 - I'm agnostic about the causes of expansion to Central/North European Russia (mostly Z282* and M558).

In South Asia I'm saying nothing specific other that certain lineages seem associated with an internal expansion from the NW, surely Neolithic in essence, as well as caste structure to some extent. All I say is that these lineages MIGHT have not just traveled in the Neolithic waves (presumably of Dravidian language) but maybe also in the Indo-Aryan one (but always from around the Indus basin, not from Europe nor Central Asia). These lineages appear in rough terms to include R1a, L and J2 especially and in this South Asian context I'd consider them original from NW India/Pakistan (even if ultimately R1a and J2 obviously stem from West Asia). There's no signal of R1a flow from Central Asia to South Asia unless you consider the southernmost area of Central Asia (i.e. Afghanistan), and in this case rather as corridor from Iran than from the North.

Repeat link for maps and other references: http://forwhattheywereweare.blogspot.com/2014/03/y-dna-r1a-spread-from-iran.html

Maju said...

"Maju, R1a-Z93 is an Indo-Iranian marker which first expanded with the proto-Indo-Iranian Andronovo culture. It arrived in South Asia well after the Neolithic in several waves, and the reason it's so common there is because of a series of massive and rapid founder effects (which is also why its STR diversity is so high there)".

David: just repeating that mantra does not make it anymore true.

It's not a problem of "STR diversity" at all: it's a problem of phylogenetic structure superimposed on the map: the "geostructure" of the haplogroup by SNP-based hierarchy, only using STR trees (and never raw diversity) to infer as much as possible origins within haplogroups for which downstream hierarchy is not yet available.

Davidski said...

The basic phylogenetic structure of R1a hasn't changed for over a year now, even after the addition of many full R1a sequences from Europe and Asia. This is how it looks.

http://imageshack.com/a/img540/9425/s4hgGC.jpg

Maju said...

Back to Krefter: " How do you explain Andronovo and its descendants being practically uniformly R1a, and mostly R1a-Z93(based on STR haplotype)?"

Because it was already there when the culture formed? It was there back then, it's there now as well... even they now speak Turkic mostly. There's no rule that says that all cultural expansions must happen because of demic replacement. A lot of cultural changes have happened without meaningful demic replacement. Even important linguistic changes! Whether those changes implied replacement or not is what we are trying to discern via genetics, but there may well not be an universal answer: Kazakhstan probably saw a good deal of replacement with the Turkic expansion but Azerbaijan or Turkey didn't. Similarly Poland may have seen dramatic upheavals back in the day but Central Asia or India not so much. Who knows?! DNA evidence is the only one who can inform us with some certainty. So I try to guide my opinion on that evidence as much as possible, even yielding to uncertainty and shrugging if the data is not clear enough. That's scientific method, I understand.

Maju said...

@David: the structure does not change surely but what Underhill in 2014 does is finding presence of those lineages that you believed in 2013 (!!!) to be European in West Asia and also finds that the uppermost branches of R1a are only found in significant amounts in Iran and Kurdistan, determining its origin with all the certainty that Occam's razor allows for.

The structure hasn't changed but the sampling bias has been substantively altered, providing us with a more clear and less Eurocentric result.

PS- Please take off the captcha: it's crazy!

Davidski said...

Maju,

The phylogenetic structure of R1a-M417 based on modern samples suggests that it's an Indo-European lineage that expanded during the Copper Age into Europe mostly as R1a-CTS4385 and R1a-Z645 (Z282+), and into Asia mostly as R1a-Z645 (Z293+).

We also have evidence of this in ancient DNA. The Andronovo and Scytho-Siberian R1a is Z282 and/or Z93. It's hard to tell, because these lineages are so closely related that their STR haplotypes are very similar.

But remains from an intrusive Bronze Age population into the western Altai are definitely Z93, and accompanied by European mtDNA haplotypes and mixed hair/eye pigmentation. The Copper Age sample from the same study belonged to Y-DNA Q and mtDNA R.

http://eurogenes.blogspot.com.au/2014/06/r1a-z93-from-bronze-age-mongolia.html

The fact that the R1a rich Tarim Basin remains are not Z93 just shows that there were multiple waves of migrations from the steppe, carrying both Z93- and Z93+ lineages.

For some reason Z93 moved south and became very successful in and around India. You'll have to come to terms with this at some point.

Unknown said...

@ Maju

"What languages those people spoke beats me. People do not only speak the language of their forefathers but actually change languages in very little time (one or a few centuries) given the right circumstances. Languages are tools and people adopt them according to their needs."

Now we're talking !

"Volga at all but rather from somewhere around Kiev. This means that, even if it is a founder effect associated to IE/Kurgan expansion, it is not necessarily something that Kurgan invaders brought from the Volga"

Something I have been stating repeatedly to deaf ears. In fact, nothing to do with Yamnaya at all "picking up", but rather a post-Cucuteni expansion.


@ Skilur
"afansevo was most likely not indo-iranian and had maybe something to do with the tocharians"

Lets not make baseless speculations on language on the sole basis of genetic hypotheses. Tocharian is not attested for 3000 years + after the said culture.

Let's analyse the DNA data on its own basis (this goes to Mr Reich) !

Unknown said...

@ Maju

". A lot of cultural changes have happened without meaningful demic replacement. Even important linguistic changes! Whether those changes implied replacement or not is what we are trying to discern via genetics, but there may well not be an universal answer."

Exactly.

David & Krefter :
Current branches of R1a might well have expanded in the copper age and later, yet I'm confident that we also will find older R1 & related branches on the steppe; which must be relate to the earliest presence of humans in north Eurasia. What we then got were periodic Y-DNA " sweeps"; the bearing of which on linguistics is often assumed through crude models and simplistic assumptions

Chad said...

I fail to understand why people throw CT around all the time. There must be a serious lack of knowledge of its ancestors and expansion. Let's not throw that name around unless you're providing substantial evidence, or any evidence at all.

Krefter said...

Mike,

We'll just have to wait for Reich's Y DNA results. As of now I lean towards Davidski's view, because I don't see any way around CWC and Andronovo Y DNA.

Unknown said...

@ Chad

" fail to understand why people throw CT around all the time. There must be a serious lack of knowledge of its ancestors and expansion."

Case in point - Chad Rohlfsen (?)

Unknown said...

Chad. The Cucuteni - Tripolye culture is not a person. It does not have 'ancestors'.
You repeatedly demonstrate your utter ignorance . All many people can do here is pontificate about ANE ratios, and have little if any knowledge on anything else. Their understanding of archaeology is infantile, exemplified by their pathetic use of 'archaeological cultures' as pseudo-historical realities which function as vectors rigidly used and forced within their pre-determined 'grand narrative'.

Chad said...

It's not just a single person. Several people throw it around. I have never seen anything that would lead one to believe that CT was influential or formed another culture after 3000BCE.

Chad said...

I won't dignify that comment with a response. When you come with data or an instance of language spread without genes, we can talk.

Krefter said...

"You repeatedly demonstrate your utter ignorance . All many people can do here is pontificate about ANE ratios, and have little if any knowledge on anything else. Their understanding of archaeology is infantile, exemplified by their pathetic use of 'archaeological cultures' as pseudo-historical realities which function as vectors rigidly used and forced within their pre-determined 'grand narrative'."

Can't wait to see your reaction to Reich's paper.

Corded ware wasn't Indo European or have anything to do with migration and ethnicity, it was a diverse nation of fluffy unicorns.

Unknown said...

Well, you should not equate you “not hearing about it” with it not being a possibility. In any case, your language suggests looking for simple answers.

Actually, as I’ve pointed out several times, the ‘dissolution’ of the C-T phenomenon had less to do with yamnaya invaders than with a diaspora-like movement of its inhabitants; moving toward the steppe, as well as the north European plain. But this was far from the all -encompassing phenomena you (incorrectly) claim for Yamnaya.

On the flip side, Yamnaya did not influence other cultures to the extent you claim. Yamnaya traits were *selectively adopted* by pastoralists eg on the Hungarian steppe who began to venture eastward as part of a new, more mobile strategy.

Afansievo did not “derive” from yamnaya, but merely has superficial resemblance to it; based on the superficial parallels in pottery etc (which are very general and not backed up by petrological studies).

Moreover, as I have already indicated in this blog, Afansievo is **earlier** than yamnaya, or at least contemoporary. It has a completely different faunal profile. Etc, etc.

I could provide references, but they’re wasted on (most) of the crowd here.

Maju said...

@David: How much aDNA from India do we have? Zero!

That there is some R1a (z93 particularly) in the steppe is not surprising. But what was the Y-DNA of IVC? My claim, and I challenge you to prove me wrong, is that it also included R1a necessarily, specifically of the Z93 type.

Sampling bias is not an argument for the directionality of genetic flow.

"For some reason Z93 moved south and became very successful in and around India. You'll have to come to terms with this at some point".

Regardless of the inexactitude of somehow making Tocharians cultural ancestors of Indo-Aryans, what they were not by any means, I don't have to come to terms with anything that goes against the evidence.

I suggest that you come to terms with Underhill 2014 first. I'm still awaiting your organized criticism beyond repetition of denial. Get the surgeon knife and do a dissection of Underhill's data, dare to do something else with that than mere denial, please. I'm demanding science, scientific criticism, not Goebbelsian repetition of an ideological fallacy.

Unknown said...

Chad

"I won't dignify that comment with a response. When you come with data or an instance of language spread without genes, we can talk"

Are you serious ?

1) Roman Empire
2) The transition from Pictish to Scots Gaelic
3) Probably Celtic
4) Greek in the Eastern mediterranean
5) Arabic (partly)
6) etc
etc


Nirjhar007 said...

@David
''The phylogenetic structure of R1a-M417 based on modern samples suggests that it's an Indo-European lineage that expanded during the Copper Age into Europe mostly as R1a-CTS4385 and R1a-Z645 (Z282+), and into Asia mostly as R1a-Z645 (Z293+).''
No aDNA from SC ASia and Iran OTH all other data suggest Indo-European to be there from Neolithic Horizon, NEXT

''We also have evidence of this in ancient DNA. The Andronovo and Scytho-Siberian R1a is Z282 and/or Z93. It's hard to tell, because these lineages are so closely related that their STR haplotypes are very similar.''
Bedouins are also of Z282* anyway really? Andronovans had Z282??? please give the data.''NEXT

''But remains from an intrusive Bronze Age population into the western Altai are definitely Z93, and accompanied by European mtDNA haplotypes and mixed hair/eye pigmentation. The Copper Age sample from the same study belonged to Y-DNA Q and mtDNA R.''
Phenotype is irrelevant anyway what makes them ''Intrusive"? Z93 is Asian!!! in E Europe it is an intrusive one.NEXT

''The fact that the R1a rich Tarim Basin remains are not Z93 just shows that there were multiple waves of migrations from the steppe, carrying both Z93- and Z93+ lineages.''
What a wishful fairy tale the actual conclusion should be that Asian aDNA also had SNP Mutations other than the Z-93 and i bet Rakhigarhi aDNA will also have Z-93- mutations.

Unknown said...

Mike, those all surely have gene flow involved. Anyone that claims Rome, Greece, or Celts didn't spread genes, is off their rocker.

Maju said...

@Mike: "In fact, nothing to do with Yamnaya at all "picking up", but rather a post-Cucuteni expansion".

There's no such thing in terms archaeological. I appreciate that we agree in other aspects but definitely not in this one.

@"I fail to understand why people throw CT around all the time".

It's pop, a fashion. It was in the news in the last years and is no doubt a fascinating culture that captures people's imagination: your charming advanced peaceful civilization, who doesn't want to dream of such "noble" roots rather than some brutal "Attila"? And it has something to do with East Europe - borderline, totally unrelated, but something.

Remember that most people discussing these matters belong to an ethno-linguistic reality that is an Indoeuropean subset, be them American, European or Asian. So it's about who they are: it's emotional. And what happens in most cases when you tell someone, who is oblivious, that his/her father is a brutish rapist slaver? Skepticism and denial. Or even when you tell them that is not even the slaver but the poor helpless victim of that criminal? Similar reaction probably. Everyone dreams of an ideal family and, sure, it's not their father but remote ancestors who we are talking about but for people with enough interest in roots and origins (many commenters surely initiated their interest looking for their own personal roots in commercial companies or otherwise but with a personal/ethnic motivation behind anyhow, it's personal!) it's almost the same.

And that explains the unjustified popularity of models like Renfrew's or the IE "paleolithic continuity" conjectures (IE panspermia I call it) or the appeal of Cucuteni-Tripolje: they offer a comforting lie to rely upon, just like Santa Claus.

But the same can be said of people who may have more sadistic tendencies (or just absorbed better the IE victor ideology) and want their ancestor to be that brutal IE "Attila" and not the enslaved victim. These people want to see the seed of those conquerors everywhere, even against all evidence. Because what's the "glory" in being the descendant of slaves? This explains other excessive and unrealistic interpretations of IE expansion and the excessive emphasis on arguing that nearly all modern lineages in their area stem from it. I call this the "Chingis Khan syndrome" because of that funny pop-science idea of the Mongol leader having caused on his own sexual prowess alone a massive genetic legacy in Central Asia. We can call this "pop-science" but, sadly enough, the root of these ideas often stems from the universities, where there are no doubt many very serious scholars but also some not so strict and too tempted by simplistic impossible "explanations" and the media limelight.

Nirjhar007 said...

David, its probably time for another post;)
Maju, How do we Know that IE folks were all of Attila/Zenghis Type??

Krefter said...

That's good insight Maju. I got interested because I wanted to be descended of Attila-types.

Usually articles online about ancient DNA or genetics, give a character and romantic side to genetics and ancient people that appeals to the authors or customers. I get tired of that, and I got a feeling we're going to see alot of it once Reich's paper comes out.

I've noticed British and American writers are the most prone to this.

Davidski said...

Here you go Maju, a dissection of Underhill et al. 2014:

http://polishgenes.blogspot.com.au/2014/03/the-story-of-r1a-academics-flounder-as.html

How you can treat seriously any paper on R1a that puts Z280 above Z282 on the R1a phylo tree is beyond me.

Nirjhar007 said...

Didn't you informed Peter about that blunder? You are a celebrated Genome Blogger aren't you?

«Oldest ‹Older   201 – 400 of 499   Newer› Newest»