search this blog

Thursday, June 18, 2015

Paul Heggarty: desperate or clueless?


Over at Diversity Linguistics Comment, Paul Heggarty of the Max Planck Institute for Evolutionary Anthropology (Leipzig) puts his foot in his mouth with a long-winded and rather whiny comment piece about two recent ancient genomics papers, Haak et al. and Allentoft et al., and the PIE question.

I don't have the time or energy right now to pick apart in detail Heggarty's ramblings, so I'll only focus on a couple of points. Firstly, here's a modified figure from Haak et al. that Heggarty put up with his post, and below that a couple of quotes with his explanation.

These data imply that Uralic-speakers too would have been part of the Yamnaya > Corded Ware movement, which was thus not exclusively Indo-European in any case. And as well as the genetics, the geography, chronology and language contact evidence also all fit with a Yamnaya > Corded Ware movement including Uralic as well as Balto-Slavic.

Both papers fail to address properly the question of the Uralic languages. And this despite — or because? — the only Uralic speakers they report rank so high among modern populations with Yamnaya ancestry. Their linguistic ancestors also have a good claim to have been involved in the Corded Ware and Yamnaya cultures, and of course the other members of the Uralic family are scattered across European Russia up to the Urals.

These are exceedingly naive and stupid comments from someone representing the Max Planck Institute. Perhaps as an ardent supporter of the Anatolian hypothesis he's feeling more than a little desperate at this point and clutching at straws? That's because anyone with even a basic grasp of European linguistics and genetics should know that:

- present-day Hungarians and Estonians speak Uralic languages, but they are of course overwhelmingly of Indo-European origin, which is easily seen in their genome-wide and uniparental DNA

- other Uralic speakers, further to the north and east, in the forest zone away from Indo-European influence, are clearly distinct from the vast majority of Indo-European speaking Europeans, because they show significant levels of recent Siberian ancestry, which was missing among the Yamnaya and Corded Ware people, and appears to be an Uralic-specific genetic signature

- therefore, it's highly unlikely that Uralic-speakers were also part of the Yamnaya > Corded Ware movement; rather, early Uralics in all likelihood began to move west across the forest zone well after the Yamnaya and related expansions from the steppe.

Heggarty also can't get over the fact that not all Indo-European speaking Europeans harbor as much Yamnaya-related ancestry as Northern and Eastern Europeans.

Above all, the Yamnaya > Corded Ware impact is much less widespread in Europe than Indo-European languages are. Much of southern Europe has spoken Indo-European languages from our earliest records (Latin and its ‘Italic’ relatives, Greek, Albanian and various other Indo-European languages of the Balkans, now extinct).

Some (low) proportions of apparent ‘Yamnaya’, ‘Corded Ware’ and north European ancestry do appear in present-day populations of southern Europe (Haak et al. 2015 Figure 3b). But such north to south population admixture is in any case expected from the historical period. The collapse of the Roman Empire and the migrations of the early medieval period were defined by major invasions and settlements of Slavic and Germanic-speaking populations into southern Europe.

The levels of Yamnaya-related admixture among present-day Southern Europeans are significant and plenty enough to explain why most of them speak Indo-European languages. All of this Yamnaya-related admixture cannot be explained by Germanic and Slavic incursions into Southern Europe during the early medieval period, because:

- most Southern European populations show very little admixture from Northern and Eastern Europe dating to this time frame (see Ralph and Coop 2013)

- R1b-M269 is the most common Y-chromosome haplogroup across much of Southern Europe, and its subclade structure among Southern Europeans, as well as the ancient DNA data from Haak et al. and Allentoft et al., suggest that the vast majority of it arrived there from somewhere in the east before the historical period but after the Neolithic.

About the only worthwhile point that Heggarty makes is that we need more ancient DNA, especially from more southerly regions, to help solve the PIE riddle once and for all.

He probably thinks that the new data will back up the Anatolian hypothesis. It won't. If Heggarty could actually understand the data from Haak et al. and Allentoft et al., he'd already know that the jig was up for his pet theory.

See also...

The ancient DNA case against the Anatolian hypothesis

Population genomics of Early Bronze Age Europe in three simple graphs

244 comments:

1 – 200 of 244   Newer›   Newest»
Nirjhar007 said...

//Both papers fail to address properly the question of the Uralic languages. And this despite — or because? — the only Uralic speakers they report rank so high among modern populations with Yamnaya ancestry. Their linguistic ancestors also have a good claim to have been involved in the Corded Ware and Yamnaya cultures, and of course the other members of the Uralic family are scattered across European Russia up to the Urals.//
I think Kristiina will give us some good suggestions on the Issue....
Another Issue is that genes and languages are not parallel to any linguistic identity, we associate them with Language on the basis modern observations....

Nirjhar007 said...

^ on the basis **of** modern observations

Davidski said...

I've got a better idea. Heggarty should pull his head out of his ass and move with the data.

Nirjhar007 said...

Its all relative David, of course the interpretation you follow for Yamnaya->CWC= Indo-European, yes it is possible but we know genes don't speak languages and one more problem is the NC European centrality of R1a1a from aDNA perspective IOW we need R1a from Yamnaya, the Karelian R1a bearer was more likely an Uralic or Non-IE speaking fella than PIE related...

Davidski said...

He wasn't Uralic, because his genome couldn't help to explain the genetic structure of modern Uralics in Europe.

And he wasn't Indo-European, because he was M417-.

Nirjhar007 said...

That why I said //the Karelian R1a bearer was more likely an Uralic or **Non-IE speaking fella**//
It had to be either of the two...

Davidski said...

Then why did you mention him if you knew he was irrelevant to the question of Indo-European expansions?

Nirjhar007 said...

oh yes i said it NOT because of the Y-DNA mutation but Geography and Time.

Davidski said...

Karlia_HG has nothing to do with the Proto-Indo-Europeans, the Corded Ware expansions, or the Proto-Uralics.

He's too far northwest, M417-, doesn't show any recent Siberian ancestry, and thus not even worth mentioning in this context.

Nirjhar007 said...

It is related R1a David yes not to R1a M-417 but surely with the growing suggestions as some has put WHG= R1a and Origin Somewhere in NC Europe...

Davidski said...

Well he obviously didn't arrive in Karelia from NC Europe, did he?

It's more likely he got there from the east as EHG groups expanded from the east across the steppe and forest steppe during the Mesolithic.

Arch Hades said...

The only way the Anatolian hypothesis could make any sense with all the info we have now about population changes in Europe is if the farmers who migrated into the steppe were like the ones who migrated into Europe. But we know they were not, they were more Georgian like and not Sardinian like. So I'd say Renfrew is pretty much dead and it's between Gimbutas and the Gamkrelidze & Ivanov school. For Renfrew to be right there would have to all these huge post Neolithic migrations and no language shift. Very unlikely IMO.

Davidski said...

Gamkrelidze & Ivanov are even more hilarious than Renfrew.

a said...

Not perfect but interesting to get a rough idea.
MDLP World-22 4-Ancestors Oracle
Yamnaya samples and CW and Sintashta both M/F.
http://www.y-str.org/p/ancient-dna.html
Kit Num: F999946

# Population Percent
1 North-East-European 51.67
2 West-Asian 26.26
3 Samoedic 9.84
4 North-European-Mesolithic 5.18
5 Indo-Iranian 3.99
6 Indian 1.92

Using 4 populations approximation:
1 West-Asian + North-East-European + Russian_North + Udmurd @ 6.405675
-------------------------------------------
Kit Num: F999942

# Population Percent
1 North-East-European 55.29
2 West-Asian 25.40
3 North-European-Mesolithic 7.29
4 Indo-Iranian 3.55
5 North-Amerind 3.40
6 Samoedic 3.05
7 Atlantic_Mediterranean_Neolithic 1.41

Using 4 populations approximation:
1 West-Asian + North-East-European + Komi + Vepsa @ 6.028195
-----------------------------------------------------------

Kit Num: M951285

# Population Percent
1 North-East-European 52.70
2 West-Asian 23.45
3 Samoedic 7.69
4 North-European-Mesolithic 7.55
5 Indo-Iranian 3.30
6 North-Amerind 3.20
7 Atlantic_Mediterranean_Neolithic 1.85

Using 4 populations approximation:
1 West-Asian + North-East-European + Mari + Vepsa @ 6.140025
---------------------------------------------
Kit Num: M020637
# Population Percent
1 North-East-European 52.63
2 West-Asian 22.05
3 Samoedic 8.95
4 Indo-Iranian 5.99
5 North-European-Mesolithic 3.33
6 Atlantic_Mediterranean_Neolithic 3.04
7 South-America_Amerind 1.38
8 Arctic-Amerind 1.34
9 North-Amerind 1.09



Using 4 populations approximation:
1 West-Asian + North-East-European + Estonian + Mansi @ 7.687467
----------------------------

Kit Num: M690970

# Population Percent
1 North-East-European 58.24
2 Atlantic_Mediterranean_Neolithic 19.97
3 West-Asian 9.67
4 North-European-Mesolithic 6.13
5 Indo-Iranian 2.71
6 Paleo-Siberian 1.44

Using 4 populations approximation:
1 North-East-European + Nogai + Swedish_V + Swedish_V @ 4.266982
----------------------------------------------------
Kit Num: F999949


# Population Percent
1 North-East-European 53.46
2 Atlantic_Mediterranean_Neolithic 25.60
3 West-Asian 10.56
4 Samoedic 3.95
5 North-European-Mesolithic 3.23
6 Indo-Iranian 2.77

Using 4 populations approximation:
1 German-North + Kosovar + Mordovian + Mordovian @ 4.225669
---------------------------------
Kit Num: F999948
Population Percent
1 North-East-European 53.94
2 Atlantic_Mediterranean_Neolithic 25.49
3 West-Asian 9.34
4 North-European-Mesolithic 6.00
5 Samoedic 1.32
6 South-America_Amerind 1.15
7 Melanesian 1.06

Using 4 populations approximation:
1 Inkeri + Kosovar + Russian_South + Ukrainian-Center @ 4.010126

Nirjhar007 said...

Dave,
1. The Oldest R1a sampled in Europe is From Karelia.
2. None of the Yamnaya samples have able to produce any R1a but R1b the Oldest Samara sample is also R1b.
3. So far the CWC culture seems to have majorly R1a.
4. Old basal clades of R1a like L-664 is also there in the NC European Area and then We have Sintashta etc though with Asian derived R1a Branches but higher levels of WHG type ancestry than ANI.
there maybe more...
The above observations gives the pattern that Samara-Yamnaya area was not the one where R1a originated at least for Europe and its clearly somewhere in North and there is also this Relation of Central Asia with CWC etc with R1a.

Nirjhar007 said...

//Gamkrelidze & Ivanov are even more hilarious than Renfrew.//
No they aren't, they have good cases of their own....

Davidski said...

You're wrong.

Nirjhar007 said...

correction: @ no. 4 its ANE not ANI:P.

Nirjhar007 said...

Don't you get what i'm trying to suggest? there is a special relation of Western Asia with Maykop-Yamnaya and there is a Special Relation of Central Asia with CWC etc of NC Europe, the first is by Hg R1b and the second is with R1a.
The Afanaseivo dna though show Yamnaya like feature but i can bet my life that when they will sample their Y-DNA they will only find R1a instead of R1b.

Matt said...

David, slightly OT from Heggarty, I was looking at the Haak et al Supplement and noticed that under its figure S7.11 and calculation of f4 (European, Corded_Ware_LN, LBK_EN, Chimp) there's what seems like a strange finding that Corded_Ware_LN is closer to LBK_EN than all Europeans are expect for Sardinian, Spanish_North, Basque, Bergamo and French South.

See here - http://i.imgur.com/7zJPVPM.png

Many of these differences are probably non-statistically significant, but that doesn't change that it seems like a strange finding to me, because if CW is the result of a population coming in and mixing with MN Europeans, and CW have much less MN European ancestry than most present day Europeans, then I would have thought they should be statistically significantly further from LBK_EN, unless something else has also happened.

I'd like to have a look at some more D stats of this form to investigate whether this also exists for other ancient farmers (EN and MN), proxies for the ancient Near East (BedouinB, Druze, etc.) and a few stats for how this behaves for Yamnaya.

D stats - http://txt.do/wfp1, http://txt.do/wfpo, http://txt.do/wfpl, http://txt.do/wfph

Any chance of you being able to run these off for me?

Mike Thomas said...

Matt

Your findings suggest that Corded Ware wasn't *the * final admixture event. Not surprising IMO
Much happened since the EBA, even if Ithere weren't massive exogenous migrations, outside certain regions like EE.

Balaji said...

Haak only made the modest claim, “These results provide support for the theory of a steppe origin of at least some of the Indo‐European languages of Europe.” The languages of Northern and Central Europe are likely to have a steppe origin. But the languages of Southern Europe may not have had an origin in the Steppe. Indeed the ADMIXTURE figure from Haak shows that Southern Europeans have distinctly more of a purple “Papuan” component at K=6 than do Northern Europeans. Moreover Davidski's f3(Pop;LBK_EN,Papuan) statistics are negative for Southern Europeans but not for Northern Europeans. I give some of them below. The statistics are also not negative for Basques and Sardinians.

LBK_EN Papuan French -0.002265 0.000638 -3.548
LBK_EN Papuan French_South 0.000876 0.000839 1.044
LBK_EN Papuan Spanish -0.003703 0.000594 -6.239
LBK_EN Papuan Spanish_North 0.001634 0.00097 1.686
LBK_EN Papuan Basque 0.00426 0.000688 6.19
LBK_EN Papuan Sardinian 0.001522 0.000643 2.366

While, I do not agree with Heggarty about Indo-European languages coming to Europe with farming from Anatolia, I think his idea that there were at least two different routes for the spread of the language family in Europe has merit.

Alexandros said...

David,

Most linguists (and old-school archaeologists) will have a really hard time interpreting all this new genetic evidence, therefore we will definitely be seeing more and more posts/comments like the one from Heggarty. Note that for obvious reasons, it is easier for a population geneticist to acquire at least some basic knowledge in archaeology/linguistics than it is for an archaeologist/linguist to acquire even basic knowledge in population genetics..

Along the same lines, I really want to see the face of certain archaeologists when new aDNA papers disprove what they believed to be facts for decades.. To be more specific, there are prominent academics of Eastern Mediterranean Archaeology, who really believe that Cyprus received minimal to none foreign admixture since the late Neolithic!! Sooner or later such beliefs will be disproved and ridiculed..

Davidski said...

Matt, here are those f3 stats.

Some of the stats didn't work out, I think because the inbreeding: YES flag needs more than one sample as one of the refs.

https://drive.google.com/file/d/0B9o3EYTdM8lQT3F0TE9OZXRHdm8/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQaFBTRjFGakNSTUU/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQTEJCem44Tk4yYVk/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQdHhVemV2QXBNLUk/view?usp=sharing

Alberto said...

@Balaji

Those results are interesting, but what do you think the Papuan element represent? I remember many stats, but in the end it was difficult to make much out of them. If it is an ASI component, shouldn't an Austro-Asiatic population from India be a better proxy?

The Papuan admixture appeared in Yamnaya, but then it dilutes quite a lot in CW and derived, and absent in modern European population (except tiny traces in Southern Europe).

Maybe that component was high in the "Armenian-like" population, and that population later had more impact in Souther Europe through Mediterranean contacts? Now we have a BA Armenian, so it could be checked somehow.

Alberto said...

@David

I think something went wrong with those stats. Shouldn't they be D-stats with Chimp as an outgroup?

Davidski said...

Whoops.

Okay, I think you're right.

M. Myllylä said...

I have no intent to argue about the origin or migration routes of IE-languanges. I however want to correct a widespread false understanding about Finnish genetic. If someone can show that I am wrong, you are welcome to show it. This is important because every evidence we have can strenghten our understanding about the history, but only if we want to see facts.

Facts:

- almost all Finnish mtdna is western or southern
- 40% of the Finnish ydna is purely western, as we understand the west during modern times
- 20% of the Finnish ydna is common only around Baltic Sea and ANCESTRAL for the Baltic N1c1
- the rest 40% of the Finnish ydna consists of various N1c1 subclades common in European side of Russia.

Most ydna groups in Finland are over 2500 years old in the Baltic Sea area, only a small proportion, around 5% is from historical time, so we are speaking about Bronze or Iron Age migrations from Neighborhood regions to Finland.

On the autosomal side the Finnish inheritance is quite heterogeneous; one one side it is clearly more western than what we see in Estonia, Latvia and Lithuania and then it is at the same time eastern similarly to Russian Uralic speaking groups.

These are facts, but when people start to mess up with genes and languages they also start to act "hot". In science we have only facts and objective methods.

Davidski said...

So far we have samples like Karelia_HG and RISE00 from near Finland. They're similar to present-day Finns, but they lack the Siberian component seen in Finns.

Thus, wouldn't you say that Uralic languages came to Finland when the Siberian component arrived there?

If so, this need not have been a migration of a pure Siberian population from Siberia, but of a mixed group from somewhere east of Finland. In any case, it looks like they arrived after the Corded Ware period, and that's all we know so far, but we'll probably know a lot more soon.

Shaikorth said...

Karelia_HG may have some weird kind of ENA considering its relation to East Eurasians, Loschbour and MA-1, but doesn't indeed explain the Siberian in Finland, or even the low level seen elsewhere in Europe showing up in Haak fits that included EHG.

Anyway, the arrival of Uralic languages to Finland is very recent according to most recent datings, so Karelia_HG does not mean Siberian did not increase in NE Europe before Uralic languages. This can be resolved with more ancient DNA from Karelia and Northern Russia between the Mesolithic and the Iron Age. Some hints are already present: the Bolshoy Oleni Ostrov site in Kola Peninsula has even more eastern mtDNA than the Karelian HG's, and predates the arrival of Uralic (Saami) to Kola Peninsula by a thousand years. The solution could be as simple as sequencing the aBOO remains' autosomes. If Siberian is there it's probably pre-Uralic in NE Europe, another matter if it's related to extinct paleo-European languages or to something else.

a said...

Another rough, 3 way comparison +/-. Very general/broad +/- range.
MDLP World-22 4-Ancestors Oracle
Kit Num: F999950-ydna Q [3214 years+/-]

# Population Percent
1 North-East-European 25.80
2 Samoedic 18.04
3 North-Siberean 9.60
4 East-Siberean 9.02
5 West-Asian 7.11
6 North-European-Mesolithic 7.02
7 Atlantic_Mediterranean_Neolithic 4.66
8 North-Amerind 3.85
9 Indo-Iranian 3.26
10 Arctic-Amerind 2.60
11 Indian 2.42
12 Mesoamerican 2.28
13 Paleo-Siberian 1.18
14 Sub-Saharian 1.03
Using 4 populations approximation:
1 Aleut + Mari + Nenets + Uzbek @ 5.978927
--------------------------------------------
Kit Num: F999942-Bateni, Russia F999942 F-T2c-4186+/- years
# Population Percent
1 North-East-European 55.29
2 West-Asian 25.40
3 North-European-Mesolithic 7.29
4 Indo-Iranian 3.55
5 North-Amerind 3.40
6 Samoedic 3.05
7 Atlantic_Mediterranean_Neolithic 1.41
Using 4 populations approximation:
1 West-Asian + North-East-European + Komi + Vepsa @ 6.028195
---------------------------------------------------------------
Kit Num: M951285-ydna R1b Yamnaya
# Population Percent
1 North-East-European 52.70
2 West-Asian 23.45
3 Samoedic 7.69
4 North-European-Mesolithic 7.55
5 Indo-Iranian 3.30
6 North-Amerind 3.20
7 Atlantic_Mediterranean_Neolithic 1.85
Using 4 populations approximation:
1 West-Asian + North-East-European + Mari + Vepsa @ 6.140025
-----------------------------------------------------------------
http://www.y-str.org/p/ancient-dna.html

M. Myllylä said...

Davidsky

"If so, this need not have been a migration of a pure Siberian population from Siberia, but of a mixed group from somewhere east of Finland. In any case, it looks like they arrived after the Corded Ware period, and that's all we know so far, but we'll probably know a lot more soon."

This is correct and proves that

- around 20-50% of the modern Finnish population brought Finnish language, which is almost same as Estonian language. Around half of that 50% (if we suggest the maximum) came from Estonia.

and

- those Finns belonging to the ancestral group of Baltic/Estonian N1c1 are either language shifters or Balts belonging to the downstrean group are language shifters.

M. Myllylä said...

Rise00 is really interesting. I am going to give extra attention for it, because it can prove something about the common N1c1 between Estonians and Finns. It is the second biggest N1c1 group in Finland.

Shaikorth said...

It's unlikely that RISE00 can show anything about relation to some baltic subclade of N1c1 because those did not exist then.

http://www.yfull.com/tree/N/

The most common ancestor of "Lithuanian" L1025+ and "Khanty-Mansi Ugrian" L1034+ was just 4300 YA. Estonian CW had unrelated Y-DNA and while it autosomally may be broadly Northeast European like, it doesn't really resemble any modern ones (even if we remove Near Eastern, Mediterranean, Siberian etc.) as K15 results show.

North Sea 37.87%
Atlantic 30.48%
Baltic 18.50%
Eastern Euro 13.16%

Haak et al said CW didn't form a clade with modern populations, it can really only tell about ancient ancestry but not recent events.

Colin Welling said...

Guys, we don't need to try to separate yamnaya like dna from uralic like dna. The two have a lot of broad ancestry in common which is normal. Of course the yamnaya PIE people were going to be genetically similar to their northern, or there a bouts, non PIE neighbors. The point of Haak and Allentoft is that an eastern (think russia) population massively changed the genetics of northern and central europe during the late neolithic in the places where IE now resides. We know this mass migration carried language with it, and that it had to be IE because no other language makes sense. Not only did they carry IE, they carried the Celtic, Germanic, and Balto-Slavic lineages. These languages only have a common linguistic ancestor deep within the IE tree. Therefore, we know that in the late neolithic a population close to PIE was in Eastern Europe.

I don't see any competing hypothesis that explains how the celtic and balto-slavic lineages migrated to their respective areas in the late neolithic from eastern europe. Seriously, please try to explain it another way and make sure to include the IE phylogeny.

Oh, its also astounding that Afanasievo, who likely spoke tocharian, which was one of the earliest PIE break offs, where like genetic replicas of the yamnaya.

Colin Welling said...

@nirj

The Afanaseivo dna though show Yamnaya like feature but i can bet my life that when they will sample their Y-DNA they will only find R1a instead of R1b.

Is that a typo? You can be sure that Afanasievo will have r1b, and will probably be dominant in r1b. But r1b is only half the IE story. Already we have ydna and autosomal evidence which strongly suggests that androvono came from eastern europe, probably the corded ware or something close by. So there you have it, two Central Asian, early, IE populations originating in eastern europe.

Nirjhar007 said...

CW,
//Afanasievo will have r1b, and will probably be dominant in r1b. //
Hogwash.
// Already we have ydna and autosomal evidence which strongly suggests that androvono came from eastern europe//
We don't know the snp of Andronovo yet, the Sintashta and Karasuk ones are Asian in Origin not E European...

Colin Welling said...

The above observations gives the pattern that Samara-Yamnaya area was not the one where R1a originated at least for Europe and its clearly somewhere in North and there is also this Relation of Central Asia with CWC etc with R1a.

I actually agree with you here.

We don't know the snp of Andronovo yet, the Sintashta and Karasuk ones are Asian in Origin not E European...

Actually they look like a mixture of ancient East European and Central/East Asia.

Colin Welling said...

Why do you think afanasievo will have r1a but not r1b?

Nirjhar007 said...

Because i'm developing this theory that Central Asia and NC Europe has a relation with R1a and Either NC Europe or C Asia is the area where R1a originated.

Nirjhar007 said...

CW,
sorry i missed this-
//Actually they look like a mixture of ancient East European and Central/East Asia.//
Autosomal basis yes but See their Y-DNA Andronovo is listed as R1a-Z-645 Ancestral to Both Z283 and Z93, so we can't put any migration on the basis of that.
The Sintashta and Karasuk ones are listed as R1a1a1b2a2 Which are SC Asian Z-94 specific clades with nothing much E European at all! yes it will be argued that they all come From European M-417 but have anyone checked yet the Early Copper Age- Bronze age samples from SC Asia or say C Asia?.

a said...

Most likely a R1b Z2103; I have not seen to many like this one. Could be quite old, and connected with Yamnaya.
ysearch,id-UZKNA from Multan, Punjab, Pakistan
UZKNA Multan, Punjab, Pakistan 12 24 14 10 11 14 12 12 12 14 13 30 15 9 9 11 11 25 15 18 29 14 15 16 18 10 11 19 23 15 12 12 11 11 13 24 10 14 12 13 10 30 24
Anyone know history for Multan?

Nirjhar007 said...

google it:D

Arch Hades said...

_"Gamkrelidze & Ivanov are even more hilarious than Renfrew."_

Why is it you think that? Can't their theory be described with what info we have now as little more than "teal people = PIE"?

Taymas said...

Hi Eurogenes Commenters,

I've never posted but I've learned a great deal from lurking here, my thanks to all commenters and especially our host. I'm really intrigued by Davidski's qpAdm result Pathan = Sintashta + Georgian + Dai and the questions that Alberto raised in response RE WHG/ANE.

In Haak, if I understood Supplementary Info 8 correctly, EHG=WHG+ANE marginally failed, but among the successful trees there were models with WHG being admixed (10% or 50% EHG input). This always seemed a strong possibility to me, because the Mammoth Steppe extended deep into Europe at times. Is there a reason to exclude this possibility? If not, how could that potentially affect the apparent discrepancy between what qpAdm is telling us versus the other tests?

I'm wondering if a relationship between WHG and EHG of that sort might have something to do with how (A) despite models showing Neolithic-Farmer + Steppe fitting extremely well for Europe, other models show Europe with too much WHG for that to be true, and (B) despite the qpAdm model fitting extremely well for Pathans, other models show Pathans having far too much ANE or Teal for that to be true. I don't understand the differences between the various tests well enough to assess this for myself.

Gioiello said...

Look at who put this sample on Ysearch and who are the closest to him (Del Badia is the reconstructed haplotype of my ancestor born in Castelfiorentino, Florence on 1449):

UZKNA Hassan Multan, Punjab, Pakistan
12 24 14 10 11 14 12 12 12 14 13 30 15 9 9 11 11 25 15 18 29 14 15 16 18 10 11 19 23 15 12 12 11 11 13 24 10 14 12 13 10 30 24
424WG Hollingsworth Armagh, Ireland
12 24 14 11 11 14 12 12 12 14 13 30 16 9 10 11 11 25 15 19 30 15 16 16 18 11 11 19 23 15 12 12 11 11 13 23 10 14 12 12 13 30 24
HFXBJ Del Badia Castelfiorentino, Italy
12 24 15 10 11 14 12 12 12 13 13 29 16 9 10 11 11 25 15 19 29 14 14 16 17 10 11 19 23 15 12 12 11 11 13 23 10 14 12 12 13 30 24
N92C5 Tognoni Italy (actually an anonymous Brazilian I think of my descent)
12 24 15 10 11 14 12 12 13 13 13 29 16 9 10 11 11 25 15 19 29 14 14 16 17 10 11 19 23 15 12 12 11 11 13 23 10 14 12 12 13 30 24



Gioiello said...

Of course there are some meaningful differences between us, thus I think that Hassan belongs to the oldest subclades after R-Z2103, those found in India as Z2106 till Z2109. I have spoken a lot about that.

a said...

For sure, mutation is old, I would wager it originates from Yamnaya time will tell.
DYS-446 is 10 not common 14.
DYS-446 mutation rate dys446 dys446 0.00095
http://www.rogersdna.com/geddna/mutate.php

Gioiello said...

DYS446=14 is the modal for the R-Z9219 and for the Balkan cluster, which may have also more. R-Z2110 has modal 13 (if you look at we European have all 13). Anyway 10 is a very low value, very unusual and may be due to a multistep mutation. Anyway these Asian samples don't demonstrate anything, because may have come to India with Indo-Europeans, i.e. I think, against my friend Nirjhar, from Eastern Europe. But the question remains the same: from East to West or from West to East? I have said many times why I think from West to East, like R1a, and next YFull results will demonstrate that.

Gioiello said...

Of course CTS9219... everything by memory...

Gioiello said...

Anyway European R-Z2109* have modal 15, thus this sample should be SNP tested. It could be also R-Z2106*... but finding the oldest haplotypes in India I think that doesn't demonstrate the origin, it could be happen also a first expansion from elsewhere. We won't know that without aDNA.

a said...

It is only the beginning. Let us see what is next. I have noticed those who were vocal, have become very quite. Others have almost
disappeared.
Next are results from Anatolia.

Matt said...

@ David and Alberto, yeah, thanks, but, yes, the chimp outgroup D stats are what I need for what I'm interested in, to try and understand the Haak f4 (Test,Corded Ware;LBK, Chimp) stat, as they should show whether Corded Ware are closer or further to a range of Neolithic Farmers than other populations, mainly present day Europeans.

Alberto said...

@Taymas

I'm wondering about this a bit more. I still don't know what it is, but something might be hiding WHG ancestry in certain populations (mostly in the Near/Middle East and S-C Asia). It might not be ANE, but maybe the Near Eastern component is inflated covering some WHG in populations that have some Arabian-like admixture. Or maybe it's something else, I still really don't know.

@David

Is it possible to get a fit for Pathans with only "extreme" populations? Even if the won't be great, but using only something like: EHG, LBK_EN and Dai (or maybe AA, whatever fits better. Maybe even including WHG). Just to get a rough idea of the proportions it gives for each.

Alberto said...

@David

Sorry, forget about the above, probably that's not going to tell us anything useful. I'll try to find a better idea.

M. Myllylä said...

Shalkorth:

"It's unlikely that RISE00 can show anything about relation to some baltic subclade of N1c1 because those did not exist then."

It doesn't need to be, because ydna is only a part of our genome and the correlation between ydna and audna depends on many historical events.

Shaikorth said...

@Mauri
Autosomal results of RISE00 are another thing, they quite likely can tell something about the history of the region. Since you have qpAdm you could try modelling modern Estonians, West/East Finns, Lithuanians, North/South Swedes, Belorussians and North/South Russians as mixes of RISE00 and something else (CEU, Chuvashes, whatever) and see if fits are good. It's just that it won't tell much about present N1c1 of the area.

M. Myllylä said...

Shaikorth,

my plan now is to do those qpAdm tests, but it takes maybe two weeks to get first results, because I want to use the data downloaded from Estonian Biocentre. It includes much more North Europeans than Haak's data. So I need newly combiled Rise-samples and preparing those samples from BAM sources takes a lot time.

Davidski said...

In regards to the Sintashta issue, and the discrepancy between modeling Pathans as part Sintashta with qpAdm and ADMIXTURE, I'm going to look into it when I put up a post about Sintashta and Andronovo.

It might take a few days though, because I want to see how the Sintashta sequences from different sources behave in qpAdm.

But I already have very little doubt that South Central Asians can be modeled as significantly of Sintashta origin, and I see no reason why there would be South Asian admixture in Sintashta. The main issue here is the extent of Sintashta ancestry in South Central Asia. It might not be as high as 70%, but I don't think it's generally lower than 50%.

Krefter said...

50%+ Sintashta ancestry in South Central Asia? That's a suprise. It's basically saying they're 50%+ East European. How can we explain their low WHG? What about mtDNA? Is R1a-Z93 really that -popular outside of India/Pakistan?

Davidski said...

By the way, these D-stats might be of interest (RISE_baBac is Swedish Battle-Axe)...

Yamnaya Corded_Ware_LN RISE_baCw Mbuti -0.0052 -1.596
RISE_baYam Corded_Ware_LN RISE_baCw Mbuti -0.0122 -3.233

Yamnaya Corded_Ware_LN RISE_baSin Mbuti -0.0024 -0.818
Yamnaya RISE_baCw RISE_baSin Mbuti -0.0014 -0.377
RISE_baYam Corded_Ware_LN RISE_baSin Mbuti -0.0084 -2.516
RISE_baYam RISE_baCw RISE_baSin Mbuti -0.011 -2.714

Yamnaya Corded_Ware_LN RISE_baBac Mbuti -0.0044 -1.381
RISE_baYam Corded_Ware_LN RISE_baBac Mbuti -0.0107 -2.991

Davidski said...

A couple more...

Corded_Ware_LN RISE_baBac RISE_baSin Mbuti 0.0059 1.488
RISE_baCw RISE_baBac RISE_baSin Mbuti 0.0054 1.146

Sintashta is a little more similar to Corded Ware from both papers than to Battle-Axe. Not really surprising.

Davidski said...

Matt,

I've got an 8GB RAM laptop and those D-stats sheets are crashing it. Maybe someone with a bigger machine can help? Or is there a way of reducing the number of D-stats to the most pertinent ones? Keep in mind that I can only run a page of D-stats at a time (something like 30), while with f3-stats I can run whole sheets without crashing.

By the way, just ran some qpAdm tests, and Sintashta can't really be modeled as the following...

Corded_Ware_LN/Pathan
RISE_baCw/Pathan
EHG/Pathan
Germany_MN/Pathan
EHG/Germany_MN/Dai

These are either failing outright or aren't very good fits probably because Sintashta lacks South Asian (Dai-like) ancestry and doesn't have as much Near Eastern ancestry as Pathans.

For example, EHG/Germany_MN/Dai is infeasible. But this works OK...

EHG 0.199
Germany_MN 0.801

Romulus said...

Wow. That was a great article by Heggarty, certainly didn't deserve the childish trashing you gave it. Calling people's ideas stupid and other ad hominem attacks are not in any way constructive, you may want to reconsider.

Davidski said...

There's nothing to reconsider. Heggarty doesn't understand the data from the two papers he reviewed.

His suggestion that Uralic people were part of the Yamnaya movement is idiotic, and so is his other claim that all of the Yamnaya-related ancestry in southern Europe might have arrived there during the early medieval period.

If you don't understand this then you're also clueless, and you might want to reconsider after taking a deeper look at the data we have.

Nirjhar007 said...

Krefter,
//50%+ Sintashta ancestry in South Central Asia? That's a suprise. It's basically saying they're 50%+ East European. How can we explain their low WHG? What about mtDNA? Is R1a-Z93 really that -popular outside of India/Pakistan?//
You can Trust on David;)...

Nirjhar007 said...

Guys,
What is the explanation of Iron age appearance of Y-DNA Q in Russian samples?.
https://genetiker.wordpress.com/2015/06/11/y-haplogroups-for-prehistoric-eurasian-genomes/

Romulus said...

@Nirjhar

The Huns?

Nirjhar007 said...

Too early to be associated with Huns i guess.

Davidski said...

These samples aren't really from Russia but from the Altai region. Q is native to this area, and it already appeared in a Copper Age sample from there.

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

So the appearance of individuals belonging to Q among the mainly R1a proto-Scythian and early Scythian groups of what is now the Russian/Mongolian border is probably due to the acculturation of local groups into Kurgan and derived cultures.

RISE492 baKarasuk R1a
RISE493 baKarasuk Q1a
RISE494 baKarasuk R1a
RISE495 baKarasuk R1a

Nirjhar007 said...

RISE 600 and RISE 601 is from which Area? Similar? I quite Agree with you David.

Davidski said...

They're almost from the same area, just a bit more southern. See map here...

http://eurogenes.blogspot.com.au/2015/06/101-ancient-eurasian-genomes-allentoft.html

Unknown said...

There is a basic problem here, of course,and it has to do with the almost total absence of neolithic aDNA from key regions south of the Caucasus. If Yamnaya was a northern extension of a surge out of the Zagros, Armenia or Iran, then that surge certainly could have entered Europe from a number of different directions,not just from the steppes. It is a simple piece of logic that whatever brought domestication to Yamnaya also brought it elsewhere. It must have occurred to some R1b types to move south of the Caucasus and Caspian long before domestication reached the steppes. If enough did then R1b stops being a mark of the steppes. And there may be other routes for it to reach western Europe.

On a side note, if this subject matter is going to aspire to be scientific, its going to become important that the over-zealousness of the steppes advocates in the past be replaced by the kind of tough self-questioning that science expects. You do need to rigorously doubt your own convictions before it becomes convincing to others. There was a "massive" migration of Germans to the American Midwest over the past 200 years, yet the language they speak is overwhelmingly English. Massive migrations may change language, but not always to the language of the migrant. And conquered sometimes end up speaking the language of the conquered. So there's room for doubt but not for jamming ideas down someone's throat.

Nirjhar007 said...

Yeah David, i got it village Verh-Uimon so the case is not much Significant,
Unknown, yes i already told that its all relative and in the absence of Written Records we can only make good guesses at best.

Davidski said...

There's always some room for doubt, and no one is saying that there isn't in this case.

The problem is that we're seeing a lot of nonsensical arguments coupled with arrogant claims and innuendos that the steppe hypothesis is somehow naive and stupid. I have no idea whether this is done due to ignorance, but in good faith, or simply to muddy the issue and confuse people.

Why would anyone ever claim that Uralics might have been part of the Late Neolithic/Early Bronze Age push into Europe from the steppe? This argument is quite simply ridiculous.

Your suggestion, "Unknown", that Yamnaya might have been a northern extension of a surge out of the Zagros, Armenia or Iran is reasonable, especially in the absence of ancient DNA from these regions, but honestly it's still a stretch considering all of the data we already have available.

Consider that most of what you need to create the Bronze Age steppe people is native to the steppe, and alien to the Neolithic farmers of Europe, who in large part were derived from farmers in the Near East. All that's missing on the steppe is some Caucasus admixture and Near Eastern mtDNA, and both of these might have easily arrived there via long distance female exogamy, which is well documented in the Bronze Age.

So I'd say things are looking very bleak for a PIE expansion from the Near East, at any time frame. Anyone who wants to still argue against the steppe hypothesis really needs to do their homework and cover all the new bases, and not just come up with a few long winded but utterly useless points to muddy the waters.

Mike Thomas said...

David.

I wouldn't think its too much of a stretch to argue that LN - EBA expansions we're seeing might not have been a monolingual affair. I mean - look at the steppe in late antiquity. A decent array of linguistic groups - Uralic (Magyars), Oghuric (Bulgars, Huns), IE (late Scythians). They were genetically all similar, most likely.

I always urge consideration of the sociolinguistic complexity in ancient groups. But as for Heggarty's specific thesis, - ive never thought the Neolithic hypothesis remotely likely.

But he is a good and credible scholar in the world of linguistics. I'd imagine it is hard to change the central tenet of one's academic life over night, especially when genetics is *still* a new field to most individuals apart from the Geneticists and avid genealogists. And even the former still get it wrong.

Davidski said...

I didn't say they weren't a monolingual affair. What I said was that they very likely didn't involve Uralic and certainly weren't responsible for the fact that today Hungarians and Estonians speak Uralic languages.

As you know, Hungarians arrived in the Carpathian Basin during the medieval period, probably from near the Volga. What relevance this has to Heggarty's doodles on the graph from Haak et al. I haven't a clue?

Using Estonians to back his case was also a mistake, since they're basically a mixture of Balts, Scandinavians and, least of all, Uralics. No wonder then that they're at the top of the graph.

We really don't need these sorts of contributions to the debate.


Mike Thomas said...

Well, yes Id definitely agree Yamnaya has nothing to do with uralic
If one were to propose an alternative to PIE it would have to be some kind of Caucasian , or perhaps extinct Pontid language.

Also in your reply to anon., it has to be realised that pastoralists groups existed South of the caucasus long before Yamnaya. ....

I vehemently anticipate aDNA from Greece, the near East and Central asia.

Gioiello said...

@ Davidski @ Nirjhar0007

I am surprised about what you wrote about hg. Q in Europe. Unfortunately you very likely haven't ever read my posts in the past: there are in Italy (but also in Western Europe) very old subclades of hg. Q (Q-M25 for instance, but also other rare subclades) which are separated from the brother haplotypes in central Asia or elsewhere from 13000 years, and may have come here in those times, as I think may have happened for hg. R1b or also R1a (even not before). Of course only a deep SNP test or the aDNA will be able to demonstrate the truth, but thinking that these samples are due to recent migrations without these tests is at least premature.

Davidski said...

I didn't write anything about hg Q in Europe.

I was talking about the hg Q in the Altai region, and among the proto-Scythians and Scythians there.

Gioiello said...

@ Davidski

Of course, but as the samples found seemed to me being Q-M25, those found in an Apennine village of central Italy in a recent paper, and we know they are sparsely in Western Europe (ask lgmayka), we should look at those old samples (but it seemed they were found also in Russia, or not?) to grasp which relation they have with the Asian and the European ones, which, as I said, are separared form at least 13000 years, i.e. if they are closer to the Asian or the European ones or interemdiate between them.

Gioiello said...

I.e. it isn't said that Scythian Q-25 in Altai have to be of "Asian" origin. They could be of "European" origin, as I think having demonstrated that Ashkenazi J and E may be of "European" origin and not necessarily of Midlle Eastern one.

Helgenes50 said...

RISE577's (Unetice) results are very close to those of the modern NW Europeans

K36

Basque 2.72%
Central_Euro 1.77%
East_Balkan 4.58%
East_Central_Euro 3.57%
Eastern_Euro 3.66%
Fennoscandian 9.21%
French 10.22%
Iberian 14.81%
Italian 5.07%
North_Atlantic 18.45%
North_Sea 21.42%
West_Med 4.53%

Davidski said...

Probably a descendent of the Bohemian Bell Beakers.

Matt said...

DavidskiOr is there a way of reducing the number of D-stats to the most pertinent ones? Keep in mind that I can only run a page of D-stats at a time (something like 30)

No problem. Really I just went big on these because I overestimated how quick and easy they were to run (so why not just throw a world and ancient picture in there too)?

I've whittled down the list to 9 much smaller sets of 30 stats (270 stats):

http://txt.do/w275 - D(Pop,Corded Ware,LBK,Chimp)
http://txt.do/w27d - D(Pop,Corded Ware,Spain_EN,Chimp)
http://txt.do/w271 - D(Pop,Corded Ware,Germany_MN,Chimp)
http://txt.do/w27t - D(Pop,Corded Ware,Spain_MN,Chimp)
http://txt.do/w27r - D(Pop,Corded Ware,BedouinB,Chimp)
http://txt.do/w279 - D(Pop,Corded Ware,Basque,Chimp)
http://txt.do/w270 - D(Pop,Corded Ware,Sardinian,Chimp)
http://txt.do/w27k - D(Pop,Corded Ware,Cypriot,Chimp)
http://txt.do/w27o - D(Pop,Corded Ware,Lithuanian,Chimp)

All together - http://txt.do/w276

Should get to what I'm looking for.

Nirjhar007 said...

David,
David,
/So I'd say things are looking very bleak for a PIE expansion from the Near East, at any time frame. Anyone who wants to still argue against the steppe hypothesis really needs to do their homework and cover all the new bases, and not just come up with a few long winded but utterly useless points to muddy the waters./
The only strong argument the Steppe Hypothesis ever had is the Horse one, the thing is that since everyone is trying to force the same hypothesis on the lens of its theories we are getting more researches trying to advocate for it which is natural also but any person who knows to observe will find its still very weak and not a single evidence has been found yet to hold it.
on aDNA a simple sampling from certain cultures of Asia will tell us how the things really are and some of them are coming soon.

Gioiello said...

Karasuk 1531-1427 BC Russia RISE493 Q1a1b1 L712+ Allentoft 2015 Geneticker
Tsagaan Asga, Mongolia TA8 Q1a2a1 Q1a2a1-L54+ Hollard 2014
Tsagaan Asga, Mongolia TA11 Q1a2a1 Q1a2a1-L54+ Hollard 2014
Tsagaan Asga, Mongolia TA12 Q1a2a1 Q1a2a1-L54+ Hollard 2014

You can see that the sample found in Russia is a subclade of Q-M25, it is found in Asia but we should see from the SNPs if it is closer to the European or to the Asian haplotypes, whereas the sample found in Mongolia belong to a different subclade:
Q-M25 Y3901 * Y3897 * L714... 58 SNPs formed 23900 ybp, TMRCA 16800 ybp
Q-M25*
Q-YP1669 14047 * YP1671 * S19575... 15 SNPs formed 16800 ybp, TMRCA 13500 ybp
id:YF01677 GBR [GB-ANT]
id:YF03186
Q-L712 Y3925 * SK1930/FGC8831/Y3912 * Y3869... 27 SNPs formed 16800 ybp, TMRCA 14400 ybp
Q-L712*
id:HG02696 PJL

Gioiello said...

Also this subclade of Q, brother to the American Q-M3, is exclusively European and Tuscan Pellegrini (I tested) belongs to it:
Q-L804 Y7555 * Y7337 * Y7344... 100 SNPs formed 15100 ybp, TMRCA 3200 ybp
Q-L804*
id:YF02368 NOR [NO-15]
id:YF03666 new
Q-Y7582 BY387/Y7606 * BY386/Y7582 * Y15697... 1 SNPs formed 3200 ybp, TMRCA 1950 ybp
Q-Y7582*
id:YF02241 SCT
Q-Y12445 PH2487/Y12446 * Y12894 * PH680/Y12445... 5 SNPs formed 1950 ybp, TMRCA 1400 ybp
Q-Y12445*
id:YF02971
id:YF03479new
Q-Y15622 Y15623 * Y15622 formed 1400 ybp, TMRCA 1150 ybp
id:YF02099
id:YF03570 GBR [GB-WKF]new
Q-Y9048 BY458/Y9050 * BY469/Y9048 * Y9051... 8 SNPs formed 3200 ybp, TMRCA 1800 ybp
Q-Y9048*
id:YF02621
Q-Y9047 Y9047 * Y9292 * Y9290... 1 SNPs formed 1800 ybp, TMRCA 1400 ybp
Q-Y9047*
id:YF02244
Q-Y10787 Y11517 * Y10787 * Y11515... 2 SNPs formed 1400 ybp, TMRCA 950 ybp
Q-Y10787*
id:YF02510 SWE [SE-Y]
id:YF02661
Q-Y15700 Y15700 formed 950 ybp, TMRCA 850 ybp
id:YF03118 SWE [SE-AC]
id:YF03485 new

Q-M3 CTS8430 * M3 * CTS7799/M884... 11 SNPs formed 15100 ybp, TMRCA 12700 y

Gioiello said...

Also this subclade of Q, brother to the American Q-M3, is exclusively European and Tuscan Pellegrini (I tested) belongs to it:
Q-L804 Y7555 * Y7337 * Y7344... 100 SNPs formed 15100 ybp, TMRCA 3200 ybp
Q-L804*
id:YF02368 NOR [NO-15]
id:YF03666 new
Q-Y7582 BY387/Y7606 * BY386/Y7582 * Y15697... 1 SNPs formed 3200 ybp, TMRCA 1950 ybp
Q-Y7582*
id:YF02241 SCT
Q-Y12445 PH2487/Y12446 * Y12894 * PH680/Y12445... 5 SNPs formed 1950 ybp, TMRCA 1400 ybp
Q-Y12445*
id:YF02971
id:YF03479new
Q-Y15622 Y15623 * Y15622 formed 1400 ybp, TMRCA 1150 ybp
id:YF02099
id:YF03570 GBR [GB-WKF]new
Q-Y9048 BY458/Y9050 * BY469/Y9048 * Y9051... 8 SNPs formed 3200 ybp, TMRCA 1800 ybp
Q-Y9048*
id:YF02621
Q-Y9047 Y9047 * Y9292 * Y9290... 1 SNPs formed 1800 ybp, TMRCA 1400 ybp
Q-Y9047*
id:YF02244
Q-Y10787 Y11517 * Y10787 * Y11515... 2 SNPs formed 1400 ybp, TMRCA 950 ybp
Q-Y10787*
id:YF02510 SWE [SE-Y]
id:YF02661
Q-Y15700 Y15700 formed 950 ybp, TMRCA 850 ybp
id:YF03118 SWE [SE-AC]
id:YF03485 new

Q-M3 CTS8430 * M3 * CTS7799/M884... 11 SNPs formed 15100 ybp, TMRCA 12700 y

Gioiello said...

Actually Pellegrini belongs to this subclade, which seems pretty exclusively European it too:
Q-L940 FGC6878/Y4803 * FGC6956/Y4823 * FGC6924/Y4820... 125 SNPs formed 18800 ybp, TMRCA 14800 ybp

Davidski said...

Matt, here you go, these should be fine, as far as I can see...

https://drive.google.com/file/d/0B9o3EYTdM8lQX09rM2VQSmFpZk0/view?usp=sharing

Alberto said...

Maybe these results that you once got with qpAdm give a hint that the Near_Eastern cluster might be doing strange things with some populations?

Yemenite_Jew
Starcevo_EN 0.396
Masai 0.137
Pathan 0.467

BedouinB
Starcevo_EN 0.335
Masai 0.150
Pathan 0.516

I hope that those samples that the Pinhasi team seem to have been working on from Turkey and the Levant come out soon. Having ancient DNA from the Near East is going to help a lot, I expect.

Alberto said...

And I quick look at the D stats Matt requested:

BedouinB Corded_Ware_LN LBK_EN Chimp -0.0374 -12.463 352644

So apparently BedouinB is not really close to ancient NE populations.

Matt said...

@David, thanks.

@Alberto:

Yes, re: the BedouinB, they seem to fit with Early Farmer clusters in ADMIXTURE to a large extent, but don't actually have that much shared drift with them. So maybe this is that there is an early farmer cluster but it's members don't sit close to one another because of high genetic diversity, or something like that? Or maybe the power of even quite low levels of African ancestry compared to different West Eurasian ancestries.

...

Here are some visualisations of some pairs of those stats:

http://i.imgur.com/eEtZ5GB.png - D(Test,CW;Spain_MN,Chimp) vs D(Test,CW;Germany_MN,Chimp)

http://i.imgur.com/8vbQYJ8.png - D(Test,CW;Spain_EN,Chimp) vs D(Test,CW;LBK_EN,Chimp)

http://i.imgur.com/TU8uHdn.png - D(Test,CW;Spain_MN,Chimp) vs D(Test,CW;Basque,Chimp)

http://i.imgur.com/NWWZSsl.png - D(Test,CW;Spain_MN,Chimp) vs D(Test,CW;Sardinian,Chimp)

The pairs above are all pretty much linearly correlated, but it seems kind of interesting that its only when you get to the modern Basques and Sardinians that North Europeans, begin to get closer than the Corded Ware.

The peak of Corded Ware increased closeness to the ancient groups seems the German Middle Neolithic, just before Corded Ware and in the same place. Visualisation here - http://i.imgur.com/KGeZO3U.png (comparing the German_MN stat to the Sardinian stat).

By those stats they're closer to Germany_MN than any modern population, while being in theory 3/4 Yamnaya and only 1/4 Germany_MN. I suppose the recent shared ancestry with Germany_MN must be *very* potent in the D stats for this to make any sense. (And it looks like there is some support for this in how tight Spain_MN and Germany_MN seem compared to even Sardinians).

Maybe this is due to post-Corded Ware drift or population movement? Possibly this helps explain why they are relying on outgroups in the Haak paper to make estimates, more than direct relatedness by f4 or D stats.

Nirjhar007 said...

Gioiello ,
Interesting but Isn't Q in Europe too rare to give any out put from there?.

Gioiello said...

@Nirjhar007

That an hg is rare, after many thousands of years, may be worth anything: see C-V20, very very rare but not found elsewhere: so it was very likely in Europe tens of thousands of years ago. After we have founder effects, genetic drift and all the rest. But you could have had a good catch from my post: if the sample of Q from Russia were linked to Indian Q rather than Europe you could bring it as a proof that IE, R1a etc came from your India.

Nirjhar007 said...

I think you are right technically:).

Gioiello said...

But there are two possibilities: or that Q came from India or was harvested from Indo-Iranian people came from Eastern Europe through India to Central Asia. We haven't any proof so far between these two possibilities... but if that Q were demonstrated to come from a subclade of Q-M25/L712 old 14400 years and present from then in Europe and not in India, it could be a strong proof in favor of an European origin.

Nirjhar007 said...

Indo-Iranian people didn't come From E Europe that you can certainly be sure of .

Gioiello said...

Unfortunately for you and all the Indian Nationalists (but didn't you say that someone said you that I am a Nationalist?) I think that all the proofs (aDNA, IE satem languages, history of the migration of Iranian peoples and before Mitanni etc.) are in favor of an origin in Eastern Europe, but I said that you rightly ask that the aDNA is tested in India or Iran, as I ask that Tyrrhenian Italy is tested for R1b. At this point we have to be subdued to "science" and not to ideology.

Alberto said...

@Matt

By looking at those stats alone, one would think that CW is very Germany_MN (and quite EEF in general) and not that much Yamnaya. Which is quite surprising knowing other data. I guess that D stats are just very sensitive to drift and/or any small admixture event.

If that's the case, a stat like D(Germany_MN,Yamnaya;Corded_Ware_LN,Chimp) should be significantly negative. Otherwise maybe CW is just more MN than we thought?

Interesting also Bell_Beaker's behaviour. They seem to sit quite close to Basques/Sardinians. Though again this might be a false impression for them being close to CW, which "seems" to be very close to MN, which in turn is close to Basques/Sardinians.

Re:BedouinB: Indeed the African admixture might be enough to explain the D stat. But why in qpAdm would they take 50% Pathan instead of more Starcevo_EN? Especially when qpAdm puts Pathans as 70% Sintashta, 18% Georgian and 12% Dai.

Whatever the case, Bedouin might not be a good choice for ancient Near Eastern cluster in admixture when we don't really know what they are exactly. But then again, we don't have any good choice yet.

Gioiello said...

Only if we think that IE languages were carried from hg J2 (which was the old idea of Dienekes, I suspect a J2 of recent Italian descent, but before likely from Asia Minor) we could think to an expansion from the Iranian plateau and older, which could explain also sanskrit and the old Indian culture...but I remember you that Tuscany has hg. J2 from probably more than 10000 years, even though I think born Southward the Caucasus.

Nirjhar007 said...

I'm a rationalist Gioiello and let me tell you that there is no reason to believe in the Steppe version of Indo-Iranian history of Course J2 is connected with IE people there is simply no doubt in that.

postneo said...

xiaohe is possibly m417 or z645. Its very far east. east of afanasievo, longitude wise its north of bangladesh.

So you have very wide spread of bronze age ancestral r1a from CW at least to the east of the Tarim basin.

to understand whats going on, apart from ancient dna from asia we need:

1) pre sintashta dna and
2) post yamnaya dna (bronze age)

Gioiello said...

The problem is that we have some haplotypes close to mine (R-Z2110*) which are found from India to Western Europe (see R-Z2106*,R-Z2109*, R-CTS7763*) and we have no explanation for that...but as I said many times India lacks the first haplotype ancestor of all (R11-L389+9) and has only the oldest R1b1-L389-, and of course lacks all the intermediate ones. The same for R1a... for that I think that these haplogroups, even come very likely from central Asia, aren't born in India, where we find only the oldest R2.

Gioiello said...

@Nirjhar007
I could give you another version of my theory:
1) IE languages brought to Tuscany from Asia Minor with J2 (J2b2?): the old Renfrew theory.
2) In Tyrrhenian Italy, where was developing hg. R from R-L51 and subclades, the Italo-Celtic branch expanded with Bell Beakers to Iberia (Atlantic facade) and to Central Europe from Southern France via Rhone/ Rhine.
3) Of course betweeen 7500 ya (migration Zilhao) and Bell Beaker, 4900ya, many things should be explained.
4) Urnfields were all hg. I like Remedello, thus once more many things said to be changed.

Nirjhar007 said...

Gioiello,
have you seen this: http://yhrd.org/tools/branch/R1a-Z2124
from the ancestral branches, it seems R1a comes from West Asia, between Iran and Turkey
the map above shows that SNP is spread also in Indonesia! Western Asia can be the area where R1a and R1b split and Central Asia where R1a started to bud, I think Uralic people had an important migration from Iran, Caspian area, as suggested from archaeology Already in the Mesolithic of the Urals has similarities with Zarzian and Caspian, and the first herders came from the south BTW i'm very sure that R1 and R2 Split happened in SC Asia or Near by.

Gioiello said...

It is always a conundrum, but step by step... first we'll solve where R-L51 were, if in Eastern Europe eastern CW or in Bell Beakers from West (or also in Middle East if aDNA found it there or elsewhere).

Nirjhar007 said...

//I could give you another version of my theory://
I think 1,2 and 4 are quite accurate Gioiello:).

Gioiello said...

@ Nirjhar007

I don't consult YHRD from when they changed format, but how is it possible?
"Y-SNP Branch Information on R1a-Z2124
The Y-SNP branch R1a-Z2124 is defined by Z2124. Additionaly, all downstream markers CTS2253, CTS6, M582, S3403, S4576, Z2122, Z2123, Z2469, Z2474 are defining this branch as well. For further information on a marker, please see marker details at Phylotree Y or consult the Y-SNP tree there.
Please note, that colours on the map only reflect a haplogroup distribution from the SNP-typed samples which were submitted to the YHRD".
How is it possible all those R-Z2124 in Japan? Neither the whole American Army could have produced them. Very likely the points on the map are the place tested for any haplotypes, and not rhe R1a found, in fact no haplotype has signed below.

Alberto said...

@Matt

On a second look, I'm even more surprised when looking at the stats with Lithuanians. In this case it's a modern population that we're comparing, so the stats shouldn't be biased by recent admixture vs. drift or post-admixture events.

We know that Lithuanians are pretty much the same as Belarusians, and very similar to, say, Norwegians.

Belarusian Corded_Ware_LN Lithuanian Chimp 0.0007 0.271 353010
Norwegian Corded_Ware_LN Lithuanian Chimp -0.0023 -0.846 353010

But in spite of CW being ancient genomes, they're equally close to Lithuanians as modern populations. But then look at this:

Germany_MN Corded_Ware_LN Lithuanian Chimp -0.0004 -0.101 327755

Here both ancient samples are on the left, so the direct and recent gene flow between them is irrelevant. And again Lithuanian is as close to both. So according to this, basically the 4 populations compared above should be almost identical??? What are we missing?

truth said...

Anyways, who cares about what this guys says, let's keep to the important things, let's not waste time with the nonsense.

Nirjhar007 said...

Gioiello, oops i gave the wrong Link, i will paste the right one which for some reason i'm unable to find at the moment:P.

rozenfag said...

Can someone explain me: soon ISABS starts, http://www.isabs.hr/index.php/welcome-to-the-9th-isabs-conference , however in their program I don't see the talk about Kumtepe. Does it mean that it will be presented on posters session?

Nirjhar007 said...

truth, can you pint out the important things one by one please?:).

Nirjhar007 said...

^**point**

Matt said...

@ Alberto

By looking at those stats alone, one would think that CW is very Germany_MN (and quite EEF in general) and not that much Yamnaya.

Yeah, if you were just looking at Corded Ware (and I expect from these stats Bell Beaker) and just looking at how close they were to the EN and MN samples relative to present day people, you'd just see at how they are closer to the EN and MN than most Europeans, and then we assume more EEF than modern North-Central Europeans.

If that's the case, a stat like D(Germany_MN,Yamnaya;Corded_Ware_LN,Chimp) should be significantly negative.

Agree. Seems like it should be negative, so would be worth a look. IRC the outgroup f3 stats from Haak et al showed Esperstedt MN (German MN) to be closer to Corded Ware than Yamnaya was. Corded Ware is also as related to EHG as Yamnaya is, under the stat D(Test,Yamnaya;Karelia_HG,Chimp).

Interesting also Bell_Beaker's behaviour. They seem to sit quite close to Basques/Sardinians. Though again this might be a false impression for them being close to CW, which "seems" to be very close to MN, which in turn is close to Basques/Sardinians.

Yes, both Bell Beaker and Basques/Sardinians share the quality of being closer to MN and EN than CW, by a similar amount, then they separate a little by Bell Beaker being further away from the present day populations who have Neolithic ancestry (e.g. http://i.imgur.com/Rr8DWqp.png) at least under this D stat with CW, and a little on the BedouinB stat.

Re: Lithuanian stats, and the Germany_MN sample, really no idea why that is like that (http://i.imgur.com/Nh8fioO.png - Lithuanian vs BedouinB).

Taymas said...

RE Davidski, Alberto, Matt

Thanks again guys, I really look forward to Davidski’s Sintashta post. Thoughts:

-As a noob, it seems to me that a big methodological discrepancy between steppe-proponents and –doubters is how literally to interpret Admixture runs. If (big if) I understand correctly, Admixture interpretation seems tricky: ancestor/descendent relationships are ambiguous and Admixture can create components from the poles of a variation-continuum when there aren’t enough un-admixed samples.

-Even if there’s no particular reason EHG-like-admixture in WHG would affect some of these results, I’d value someone's more knowledgeable take on that issue.

-We’ve now got Haak/Allentoft models, Admixture, etc with Corded/Sintashta being heavy Yamnaya, but qpAdm and D-stats saying heavy GMN. I’m very confused. If I understand correctly, Alberto was asking if it’s a matter of our components (EEF, Teal, and I'd wonder even WHG) being mixtures themselves, and Matt is proposing that the discrepancy could be various tests’ different reactions to time-since-common-ancestry. Am I keeping things fairly straight?

-Davidski, I humbly offer that I think it'd be really cool to see an update of your PCA from the Teal post with more ancient genomes and Admixture components.

Matt said...

Taymas: Matt is proposing that the discrepancy could be various tests’ different reactions to time-since-common-ancestry

Sort of. My thinking is more questioning why Corded Ware seem to be closer to EEF (LBK_EN, Spain_EN and also more markedly Germany_MN) on the f3, f4 (in the Haak paper) and D-stats (from Davidski) than many modern populations which are supposed according to qpAdm to have much more EEF ancestry.

I was thinking one possibility is time could lead to this. I don't really understand how insensitive D stats are to time and drift though.

Another possibility is what was mentioned in the Haak paper that (and which I think David has mentioned before as well):

"f4(Corded_Ware_LN, Yamnaya; Karelia, Chimp) = -0.00001 (Z=0.0). If Corded_Ware_LN was a simple mixture of a population related to our Yamnaya sample and of Neolithic Europeans, this statistic should be negative. However, if Corded_Ware_LN is descended from a population that has a higher proportion of EHG ancestry than the Yamnaya population, then the dilution of EHG ancestry due to European Neolithic admixture (which would cause the statistic to be negative), would be counterbalanced by its increase due to this higher EHG ancestry (which would cause it to be positive)"

So, if Corded Ware is more EHG+MN, then it would need less EHG than it is modelled with Yamnaya, and so have more MN, and might then fit better as being more Early Neolithic than various modern populations, and that might come closer to resolving what I see as a bit of a discrepancy. That wouldn't contradict a steppe model of Indo-European exactly either.

But it would need quite a boost of MN from this to effect to explain that pattern. And it still seems like that wouldn't be enough on its own without a very big change, so then you might need some other effect like differences in Neolithic farmer ancestry flowing around after during and after Bronze Age as well and probably some additional WHG / African / East Asian ancestry in the right places at the right time.

Matt said...

Also, Taymas, with qpAdm modelling, although I might be simplifying a little, I think that essentially what it does is fit a population A as a combination of populations B, C, D, etc. (let's call them Set 1) based on how they relate differently to a bunch of other populations (let's call them Set 2) under f4 statistics. (Either set 1 and set 2 is called right or left in the software, can't remember which).

For instance, let's say a Pop A has a lot of relatedness to Native Americans, but relatively low relatedness to East Asians, and Yamnaya, EEF and WHG are in Set 1, and Native Americans and East Asians are in Set 2. qpAdm will give Pop A a lot of Yamnaya ancestry. This is because Yamnaya is the closest to fit the bill of having a lot of relatedness to Native Americans and quite low to East Asians, compared to EEF which is lower to both and WHG which is higher to at least East Asian.

The idea is that using a lot of these Set 2 populations to cross check will converge on the real mix via statistics by fitting as many of these relationships as once. The combination is a kind of "weighted average".

This does mean though (and this is mentioned in the Haak supplement), in theory a problem with this can arise when you have two populations who could go in Set 1 who it would be hard to tell apart from their relationships with the Set 2 populations. Like if they're a clade compared to all world populations. Or it may be relatively hard to tell a blend of EEF+EHG from Yamnaya, based on how related each would to be populations outside Europe (like the Magic Set used in Haak).

So I think it is worth direct cross checking qpAdm with how related a population that models as something is to the population it is supposed to model as. Like if Corded Ware models as having lots of Yamnaya ancestry compared to EN, but seems closer to EN than some populations who are supposed to have more EN and less Yamnaya according to qpAdm, then how does that happen.

Some of this may end up applying to trying to model populations with Sintashta as well, I guess.

@ Davidski, btw a few more sets of D stats if you don't mind (np if you're busy or anything),

http://txt.do/w2y7 - D(Test,Corded Ware;Yamnaya Chimp)
http://txt.do/w2y2 - D(Test,Corded Ware;Iraqi Jew Chimp)
http://txt.do/w2yv - D(Test,Yamnaya;Corded Ware Chimp)

Mike Thomas said...

Matt

"So, if Corded Ware is more EHG+MN, then it would need less EHG than it is modelled with Yamnaya, and so have more MN, .."


So could this have something to do with CWC having straight up EHGs guys (eg from around the Baltic) mixing with Central European MNs?

VOX said...

From the paper (Allentoft et al.):

"Intriguingly, individuals of the Bronze Age Okunevo culture from the Sayano-Altairegion (Fig. 1) are related to present-day Native Americans (ExtendedData Fig. 2d), which confirms previous craniometric studies. This finding implies that Okunevo could represent a remnant population related to the Upper Palaeolithic Mal’ta hunter-gatherer population from Lake Baikal that contributed genetic material to Native Americans."

If you look at Extended data figure 5 you’d see that South Asians have a lower fst with this population compared to others. I’m confused how this population could have near intercontinental distances with East Asians and Yamanaya related populations, yet be so close to South Asians. Does anyone have an explanation?

Tobus said...

@Matt/David:

I did some testing with the number of lines in the DStat input file and found:
16 lines: 4m 13s
32 lines: 5m 04s
64 lines: 5m 54s
96 lines: killed after 1hr 10m

So it seems there's a sweet spot for maximum efficiency.

I was testing with your latest requests Matt, so here are the results: https://drive.google.com/file/d/0B1Z5v96nWWmhYnFDczRUVXRhVTQ/view?usp=sharing. Note that my data is slightly different to David's so I had to change "Germany_MN" to Baalberge_MN and "Ukrainian_West" to "Ukrainian".

saman sistani said...

On Dienekes' website he has linked to a study which states it has become possible to harvest and sequence ancient DNA from hotter environments, which makes it more possible now to sequence middle eatern ancient DNA. In Iran's shahreh-e-sukhteh, which is about 5000 years old, there is a graveyard of upto 40000 skeletons, I wonder if they can salvage some DNA from there.

Davidski said...

That study is now somewhat irrelevant to the Middle East, because quite a few ancient Middle Eastern samples have already been sequenced. The results will be published soon, probably this year.

Balaji said...

Alberto,

You wrote in reference to the purple component at K=6 in the Haak ADMIXTURE figure, “Maybe that component was high in the "Armenian-like" population, and that population later had more impact in Southern Europe through Mediterranean contacts.”

The commenter “Unknown” has a similar opinion.

I believe that something like that happened. The Near-Eastern ancestors of the Yamnaya must have had two times the amount of this component that Yamnaya have since EHG have none of it. The Near-Eastern ancestors of the Yanmanya must have been the ones who taught Yamnaya how to herd cattle and how to work metal, This must have done by been men and not just the “farmer wives” that Davidski talks about. The Near-Easterners must also have imparted their language to the Yamnaya.

These same Near-Easterners must have also moved along the Mediterranean to Southern Europe spreading ANE, the purple component and Indo-European languages. In other words, Indo-European languages reached Europe by a Northern route via the Steppe and a Southern route via the Mediterranean.

Alberto said...

@VOX

The Fst distances of Okunevo in Extended data Figure 5 are certainly interesting. I don't think we still understand well the relationship between Siberian population and ANI populations. But there must exit something.

Note that Okunevo is far from WHG and EEF. Also not too close to East or SE Asian populations like Dai, Kinh, Han or Japanese. And it's closest to Indian populations, including South Indian Dravidian speakers like Tamil or Telugu.

Interestingly, while not being close to any ancient European population, including Yamnaya, CW or Bell Beaker, (and also BA Armenian), it's closer to modern Europeans like Finns, NW Europeans and Iberians.

It must all have a meaning, but I'm afraid we still don't have a good clue about it.

Alberto said...

@VOX

Forgot to add that in any case, the distances between Okunevo and even the closest population (Punjabi) is still very high. So we have to put those relationships into perspective.

@Balaji

Yes, the hypothesis is perfectly possible, and we just have to wait for more samples from relevant places to know if that's indeed what happened or not.

A bit unrelated question (though still related to the Papuan component). Do you have a good idea about the difference between ASI and ANI? As far as I understand, ASI should be present in Papuan, Dai and in general in SE Asians. Among Indians, probably highest among Austroasiatic speakers. But what about ANI? Is it related to Dravidians or something else? And where/how did it originate? Do we know about it?

Mike Thomas said...

@ Balaji

Thanks for your thoughtful comments

you say 'The Near-Eastern ancestors of the Yanmanya must have been the ones who taught Yamnaya how to herd cattle and how to work metal, This must have done by been men and not just the “farmer wives” that Davidski talks about"

First of all, lets be specific. It was in Majkop that the earliest developments occurred, not Yamnaya. Secondly, I wouldn't make generalized statements about Majkop metallurgy without empirical evidence. No doubt influences came from the south. But the exact nature of Majkop metallurgy is till not known. We have an oxymoron; massive amounts of metallurigcal deposits in chieftains graves and little has so far been found in terms of actual worked mines and workshops. So perhaps those big Majkop chiefs were traders? Whatever the case, they show clear material culture differences (eg settlement types, metal artefacts) to groups south of the Caucasus, despite some similarities, such as pastoral economies and the raising of kurgans.

'The Near-Easterners must also have imparted their language to the Yamnaya"

Possibly. But there is no "must" in socio-linguistics....


"In other words, Indo-European languages reached Europe by a Northern route via the Steppe and a Southern route via the Mediterranean"


Quite possibly. But I'd contend it was even more complex than that

Tesmos said...

Davidski there is something strange with Sintashta Eurogenes K15 results. That genome has around 40% North Sea do you have any ideas why?

Davidski said...

The K15 isn't designed to test ancient samples at such a fine scale, because it's based on very recent drift. You can only use it to compare ancient and modern samples if they're from distinct populations. Even then I wouldn't really try it unless for a bit of fun.

Matt said...

@ Mike Thomas - So could this have something to do with CWC having straight up EHGs guys (eg from around the Baltic) mixing with Central European MNs?- I don't know if EHG per se was in the Baltic rather than only Russia, still I think it could do. Although not totally EHG necessarily, as I think the Haak paper suggests that an intermediate level of EHG between the Samara Yamnaya and EHG might work better. Or even the Western Yamnaya having some straight up mixing between EHGs+ENs maybe?

@ Tobus thanks.

It looks like the D(Test,Corded_Ware;Iraqi_Jew,Chimp) stats again find that modern Europeans are generally less related to Iraqi_Jew than Corded Ware is, or not significantly different.

Yamnaya is also significantly less related to Iraqi Jew than Corded Ware is, despite the Haak model supposed to be modelling Yamnaya as having input that resembles Iraqi Jew or Armenian. So Yamnaya continues its trend of being remote from even populations that are supposed to be a proxy for those that contributed to it.

Quick plots - http://i.imgur.com/Xg8lTCX.png (LBK / Iraqi_Jew) / http://i.imgur.com/Fi8If3Z.png (Sardinian / Iraqi_Jew) and a plot comparing D(Test,Yamnaya;CW,Chimp) to (Test,CW;Yamnaya,Chimp) - http://i.imgur.com/SqKhZIb.png.

Also Alberto from the stats run off by Tobus

D(Baalberge_MN, Yamnaya; Corded_Ware_LN, Chimp) -0.0094, Z = -1.723
D(Spain_EN, Yamnaya; Corded_Ware_LN, Chimp) -0.017, Z= -4.737
D(Starcevo_EN, Yamnaya; Corded_Ware_LN, Chimp) -0.0255, Z = -3.218
D(SwedenSkoglund_MN, Yamnaya; Corded_Ware_LN, Chimp) -0.0206, Z = -3.777

So Baalberge_MN is actually under these stats less close to CW_LN than Yamnaya is, despite CW_LN supposedly being closer to Germany_MN than modern day Europeans are in other stats...

Mike Thomas said...

@ Matt


"Or even the Western Yamnaya having some straight up mixing between EHGs+ENs maybe?"

That's what I had always suspected on a theoretical level. But, of course, we need direct data. We need a full Mesolithic -> Iron Age data set to really understand what happened in any given region.

Alberto said...

@Matt

"So Baalberge_MN is actually under these stats less close to CW_LN than Yamnaya is, despite CW_LN supposedly being closer to Germany_MN than modern day Europeans are in other stats..."

Yes, but if indeed CW is 3/4 Yamnaya and 1/4 Germany_MN, that was to be expected. The strangeness was in the other stats, showing CW very Germany_MN-like.

I guess it's what you said above, the fact that Germany_MN was the exact population (in space and time) that contributed to the CW samples (that are also from Germany) has a big impact in D stats. Here in direct comparison:

D(Baalberge_MN, Yamnaya; Corded_Ware_LN, Chimp) -0.0094, Z = -1.723

The difference is not too significant. Probably because the Yamnaya samples are from Samara, and so not the exact population that contributed to German CW.

I find interesting also that Yamnaya, as you said, looks very distant from populations that might have contributed to it. Not just Iraqi_Jew, but also Georgian or even Adygei (Adygea being North Caucasus -capital Maykop- and speculated to have steppe ancestry). This suggests (always caution is recommended when interpreting D stats) two things:

- Whichever population contributed to Yamnaya from the south, was probably very distant from EHGs (a bit surprising if they shared the high ANE plus some WHG ancestry).
- Gene flow went south to north, but there's not much evidence for north to south gene flow (Yamnaya back to Caucasus and Near East, in this case).


@Mike

Indeed. We're still missing some important pieces of information. From Western Yamnaya to the pre-CW Baltic we don't know exactly how those populations were and how they might have contributed to the LN/BA Europeans.

Tesmos said...

David, i see will you create a new calculator based on these new samples?

Grey said...

@a

"Most likely a R1b Z2103; I have not seen to many like this one. Could be quite old, and connected with Yamnaya.
ysearch,id-UZKNA from Multan, Punjab, Pakistan"

Interesting. Very old city, natural defensive position, on main route into India - sounds like exactly the sort of place I'd expect to find (small) amounts of ancient R1b.

Grey said...

VOX

" I’m confused how this population could have near intercontinental distances with East Asians and Yamanaya related populations, yet be so close to South Asians. Does anyone have an explanation?"

I have no opinion on the specific case but logically if you imagine a cross shape made up of five squares and people from the southern square move into the center and top squares and then later people from the east and west squares expand and meet in the center then the north square and south square population become cut off from each other.

I imagine things like that happened a lot.

VOX said...

Hi Alberto and Grey, good feedback and commentary. In the future it would be interesting to see this MA1 related population included in a tree-mix diagram among WHGs and EHGs. I'm curious to see how they would look like in K8 as well.

Chad Rohlfsen said...

David,

I have been looking at the Okunevo, and I wonder they aren't basically intermediary between EHG and MA1, with maybe some Siberian. It's hard to tell if they really have any Caucasus admixture. Could you try the following Dstats?

Mbuti Karitiana Karelia_HG RISE_baOku
Mbuti MA1 Karelia_HG RISE_baOku
Mbuti Yamnaya Karelia_HG RISE_baOku
Mbuti Karelia_HG MA1 RISE_baOku
Mbuti Ami Karelia_HG RISE_baOku
Mbuti Loschbour Karelia_HG RISE_baOku

Chad Rohlfsen said...

Sorry, also this one

Mbuti Ami Karitiana RISE_baOku

Grey said...

Interesting (to me) google find

http://www.buber.net/Basque/Euskara/Larry/metal.names.html

The Basque language has native words for iron, gold, silver etc but loan words for copper, tin and bronze.

" `tin', `copper', and `bronze'"

Grey said...

eztainu:`tin'
kobre: `copper'
brontze: `bronze'

Krefter said...

It looks like Felix has that Neolithic DNA from Neolithic Turkey!!!!! He has it on his list of Pending DNA files to analysis.

http://www.y-str.org/p/ancient-dna.html

Davidski said...

Yeah, and from Late Neolithic Turkmenistan.

It looks like we're about to find out a lot of what we've been wondering about.

Mike Thomas said...

Ha ha crunch time !

Krefter said...

One is 10,000 years old and from the Jordan.

VOX said...

Quite interested to see if some of the Near Eastern samples would be similar to Bedouin or not! BTW Davidski curious to know what Linux distro yo do your data crunching in, if not Mac.

Chad Rohlfsen said...

Hmm.. just playing around with the Okunevo. I can model the Yamnaya as 55% Okunevo, 45% WHG. Very odd and likely off. David, do you have time to check those Dstats?

VOX said...

Hi Chad,

Doesn't that prove difficult since the Yamnaya has some near eastern caucasus like admixture? How about the EHGs? Can they be modeled as a Okunevo/WHG mix?

Davidski said...

Chad,

Mbuti Karitiana Karelia_HG RISE_baOku 0.0127 2.047 167023
Mbuti MA1 Karelia_HG RISE_baOku -0.0463 -5.159 121554
Mbuti Yamnaya Karelia_HG RISE_baOku -0.0623 -11.617 166906
Mbuti Karelia_HG MA1 RISE_baOku -0.0327 -3.579 121554
Mbuti Ami Karelia_HG RISE_baOku 0.0407 7.32 167023
Mbuti Loschbour Karelia_HG RISE_baOku -0.0791 -10.598 165296
Mbuti Ami Karitiana RISE_baOku -0.0542 -11.668 171026

VOX,

I use Ubuntu for everything, but I had to download a few extras.

VOX said...

I duel boot with Linux Mint but on another computer I have Manjaro which refreshes with new software every update.

Things are also looking good for Ubuntu since they will have "Snappy Package" manager which isolates apps and prevents breakages allowing for stable system upgrades rather than reinstall. They also have the phone/desktop OS convergence concept going for them. Looking forward to Mid 2016 to see how it all turns out.

Tobus said...

David, where did you get the Okunevo sample from and was it easy to get it into the same format/files as the Haak samples? I wouldn't mind trying to run some tests of my own if it's not too hard to get it set up.

Nirjhar007 said...

David,
//Yeah, and from Late Neolithic Turkmenistan.//
Superb! from 5-4.5 KYBP though we don't know Female or Male:P but At Least Autosome will be very interesting!!....
There should be R1a-M417.....
But From Which Study? what do you know??.

Gioiello said...

" Grey said... Interesting (to me) google find http://www.buber.net/Basque/Euskara/Larry/metal.names.html
The Basque language has native words for iron, gold, silver etc but loan words for copper, tin and bronze.
" `tin', `copper', and `bronze'"
eztainu:`tin'
kobre: `copper'
brontze: `bronze' "

These words are from Latin, or better from some Neolatin languages, very likely Spanish, and don't demonstrate anything.

Gioiello said...

@ Grey

Very astonished that a great scholar of Basque says about the word for "iron": "Exceptionally interesting is . Since there is good evidence that the Celts introduced iron into the Basque Country, we might have expected Basque to borrow a Celtic name for the metal, but that didn't happen. The comparative evidence makes it pretty clear that the earliest form of the Basque word was *, or just possibly *, which seems less likely but cannot be ruled out. And it is not so easy to connect this with `blue' [...]".
It is clearly linked with Latin "ferrum" from *fersom, and many Middle Eastern languages like Hebrew "brzl" from some language of the places where iron was firstly used. Thus a wanderwort.

Davidski said...

Tobus,

Get in touch with Martin Sikora and request the haploid dataset in PLINK format. Then use Convertf and PLINK to convert and merge the datasets.

Or you can download the diploid genomes from Felix' site and merge them with your Haak et al. dataset...

http://www.y-str.org/2015/06/101-ancient-eurasian-dna.html

Nirjhar,

I can't imagine R1a-M417 in Turkestan 5000-4500 years ago. Maybe, just maybe, it was beginning to move out from Europe into Asia along the steppe at that time.

Ramber said...

Out of topic: Can I ask you something regarding West Eurasia K8 admixture. I notice that in K8 admixture spreadsheet, the Mongolian samples have just slightly less Caucasoid admixture than Altaian and Khakass samples around 12-13% Caucasoid. I also notice that the Mongolian samples have sort of similar amounts of Near East and WHG to the Altaian and Khakass. However there is also another population sampled called "Mongola". This one seem to have about 8-9% less Caucasoid (Near East-WHG) compared to the Mongolian samples. This makes me think that the Mongolian samples are actually Turkic tribes who become Mongolized. I heard somewhere that the western and parts of central Mongolia used to inhabited by many Turkic tribes before Genghis Khan comes to power while the Mongol tribes come from the Eastern part near Manchuria. I think that many of these Turks eventually got assimilated and become part of Mongol ethnic group. So it is likely that the real Mongols are actually the Mongola sample compared to the Mongolian sample who are likely to be originally Turkic tribes. What do you think about this?

Davidski said...

The Mongola sample comes from the HGDP and you should be able to find some info online on where they were sampled. I think it was northern China.

The Mongolians are from "The genome-wide structure of the Jewish people" by Behar et al. They're probably from Mongolia, but you should check that. Their sampling location might be in the study.

Petter Johansson said...

I think it is hard to make a model of all Uralic people spreading a Siberian component, since Mordovian people at the very Urheimat have so little Siberian and Mari have gotten it mainly from Turkic people. Western Finns and Estonians also have very little. Like other people suggest above, I think it is mainly an Arctic European thing, Siberian genes have been spread along the Arctic coast. Of course, Uralic peoples taking a very Northern route, such as Saami, might have helped spread it.

In any case, this does not speak against the PIE steppe model. Uralic and Indo-European peoples may simply have spread the same genes, with some small differences. Their Urheimats are adjacent to each other, after all, so it is only expected...

capra internetensis said...

Sweet! Okunevo is like Amerindian with less East Asian (but still has a fair bit). No Y DNA, mtDNA A8 and H6.

Davidski said...

Petter,

All Uralic-speaking people, except most Hungarians, carry the Siberian component. Even most Estonians and all western Finns have it, and Mordovians have a fair bit of it, even though they're probably largely of eastern Baltic origin.

It's obviously an Uralic-specific marker, even if it wasn't the main component of the early Uralics.

The fact that the Proto-Indo-Europeans and Proto-Uralics were neighbours is another argument against Heggarty's ramblings, but it doesn't change what I said above.

Alexandros said...

Really looking forward for the upcoming Neolithic Anatolian sample! Does anyone know, is this a male or a female?? If male, his Y-hg will be revealing! I beleive there is a high likelihood of finding G2a people in that region in the specific era. Agriculture spread westwards from SE Anatolia (North Levant) and probably the region of NW Anatolia (from where the specific sample is derived) was a gateway to the difussion of agriculture to the Balkans. What we don't know however, is whether European HGs (mainly bearing I2) also resided in that region, as it is very close to continental Europe. My bet for Y-hgs therefore would be G2/I2, but also open to other possibilities (E-V13, T, J2?)..

Davidski said...

Unfortunately, the sample from Barci is a female.

Turkey Bar100 Pottery Neolithic 8400 ybp Female 237331 SNPs

Alexandros said...

Pitty.. All this hypothesizing for nothing.. Still looking forward for the autosomal data!

Ramber said...

But if Mongola and Mongolian have such different amount of admixture, should Mongols actually be separated genetically into two ethnic groups? I was surprised by the big difference between Mongola and Mongolian and how the Mongolian samples have similar genetic composition to Turkic people like Altaians and Khakass more than their eastern neighbors (Daur, Oroqen, for example).

Ramber said...

That is the reason why I think it is likely that the Mongolians sampled are actually Mongolized Turkic peoples while the real Mongols are actually the "Mongola" sample. I heard that the Mongols actually come from eastern part of modern day Mongolia, near Manchuria, before they start to spread and conquer Western and Central parts of the country.

Ramber said...

That is the reason why I think it is likely that the Mongolians sampled are actually Mongolized Turkic peoples while the real Mongols are actually the "Mongola" sample. I heard that the Mongols actually come from eastern part of modern day Mongolia, near Manchuria, before they start to spread and conquer Western and Central parts of the country.

Davidski said...

The Mongolian sample might be from the Altai region, where there was a lot of mixing going on since the Bronze Age.

The Mongola sample is probably from China near the southeastern part of Mongolia, where there should be less European/West Asian influence.

Petter Johansson said...

Davidski wrote:
"All Uralic-speaking people, except most Hungarians, carry the Siberian component. Even most Estonians and all western Finns have it, and Mordovians have a fair bit of it, even though they're probably largely of eastern Baltic origin.

It's obviously an Uralic-specific marker, even if it wasn't the main component of the early Uralics."

Probably, but the fact still remains that the Siberian component is higher in the Arctic region, which is further away from the Uralic Urheimat, where the Siberian component is much lower. This leads me to believe that the Siberian gradient in North-Eastern Europe is mainly the result of a more ancient Siberian spread to the region, rather than a recent Uralic migration. The original Uralic speakers where then similar to Indo-Europeans, albeit missing part of the farmer genes and carrying a minor Siberian component, probably the result of living adjacent to Siberian peoples.

Some people of course also speculate the Pre-Proto-Uralic peoples where Siberian before arriving at their Urheimat. I consider this neither right nor wrong, because it is just speculation, the linguistic evidence is weak by all standards.

Alberto said...

Those samples listed by Felix come from this study that I think someone linked to already:

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0129102

Apparently the sample from Turkey should be very high coverage. If I'm not mistaken, usually from ancient remains they typically get <1% endogenous DNA. But here by using parts from the petrous part of the temporal bone they managed to obtain ~50% in several samples (the one from Turkey, the ones from Hungary and the one from Serbia).

Unfortunately, it didn't work so good for others from hot (arid or humid) weather. The 10.000 ya sample from Jordan yielded some 1% which might be good to get enough data, but the one from Turkmenistan was the one with lowest endogenous DNA at 0.1%, so I'm not sure how good that one might be.

All very interesting, can't wait to see them.

Ramber said...

Thank you very much. So the Mongols from Eastern Mongolia should have less European/Near East than those from Western Mongolia based on this K8 spreadsheet? I was surprised to see the big difference with Mongolians being like 12-13% European/Near East while their eastern relatives only 3-4% European/Near East.

Petter Johansson said...

There is also another things which makes me wonder wether Siberian is truly a Uralic marker, and that is the fact that Siberian mixture is different in different Uralic peoples. Also, Turkic migrations to the Uralic Urheimat are historically attested, and must have had an effect on Mari and Mordovian genes. In short, the model of a Siberian admixture spreading through Uralic migrations is difficult to say the least.

Davidski said...

Petter,

Discussing components from ADMIXTURE is only useful up to a point. Have a look at the D-stats at this link...

http://eurogenes.blogspot.com.au/2015/05/the-lneba-like-one-big-party.html

You can see that the Uralic speakers, apart from Hungarians and Estonians, show similar patterns that are clearly distinct from those of the North European Indo-European speakers. This is because they have relatively high levels of ancestry that the Indo-Europeans lack, or at least have a lot less of it.

You can argue that Finns behave in this way because they have Saami admixture, and Saamis have Arctic admixture, while the Mordovians have Turkic admixture, and so on and so forth. But I won't be convinced, because I can see some very strong correlations there between language and genes.

Ramber,

The Mongola are from China. I have no idea how eastern Mongolians would come out in the K8.

Balaji said...

Alberto,

The Reich Lab people were the ones who introduced the concept of ANI and ASI. In their latest paper two years ago, the estimated the proportion of ANI as f4(YRI,Basque;India_pop,Onge)/f4(YRI,Basque;Georgian,Onge). ANI is basically the ancestry that is shared with West Eurasians and ASI is the rest. Since there is not much UHG/WHG, ANI=BEA+ANE, In the Moorjani et al. paper (Nature, Sept 5, 2013), the population with the highest ANI was Pathans with 70% and the one with the lowest ANI was Paniya with 17%. The Papuan component is higher in populations with less ANI.

Mike Thomas,

You are right that Yamnaya probably did not really master metallurgy but instead got their metal tools by trade. Their descendants, the Corded Ware and Battle Axe people had to revert to stone tools. Bronze technology in Europe must have arrived by the Southern route and spread North from there along with people leading to a waning of the Corded Ware and Yamnaya influence that has been documented by Haak. In this sense, Heggarty may be right in asserting that the Steppe migration had only a limited effect on Europe.

Matt said...

Alberto: - Whichever population contributed to Yamnaya from the south, was probably very distant from EHGs (a bit surprising if they shared the high ANE plus some WHG ancestry).

Trying to sum up the lines of evidence we have about the non-EHG ancestry in the Yamnaya

1) Yamnaya have reduced affinity to Native Americans and MA1 relative to EHG; so the new population must have been less "ANE" related in a sense. And Yamnaya has reduced affinity to EHG itself.

The stats for this are f4(Karelia_HG, Yamnaya; MA1, Chimp) D = 0.00436, Z = 6.7 and f4(Karelia_HG, Yamnaya; Mixe, Chimp) D= 0.00401, Z= 8.3, and f4(Karelia_HG, Yamnaya; Samara_HG, Chimp) D= 0.00594, Z= 11.3

2) The lowest stat of the form f4(Karelia_HG, Yamnaya; Test, Chimp) are with Iraqi Jews, at D = -0.00176, and Z = -4.2; this suggests Iraqi Jews are the most alike to the "other Yamnaya ancestry" population.

But this stat isn't that strong (and e.g. the same stat for Saudi D = -0.00137 Z = -3.3), so the Yamnaya population isn't that much closer to Iraqi Jews than Karelia_HG (even though it is quite more distant from EHG than other EHG is).

3) qpAdm fits the outgroup relationships for Yamnaya decently well with Karelia_HG at around 50% and South Caucasus / Iraqi Jews at around 50%; so whatever the extra ancestry is in Yamnaya, it works like 50% Iraqi Jew to change relatedness to Native Americans, East Asians, Africans, etc.

Then you have these D stats of the form D(Test,Corded Ware,Test2,Chimp) showing Corded Ware is actually much closer to virtually all ancient and modern "Neolithic" / Middle Eastern populations than Yamnaya. Even Iraqi Jews. Particularly Corded Ware is as close or closer to EN / MN Europe as modern North-Central Europeans.

At the same time, Corded Ware is also actually closer to Ust Ishim (and I think Dai as well?) than Yamnaya, so should be more "Crown Eurasian".

So the reduced affinity to EHG and outgroup modelling indicates a Near Eastern contribution, but present day Near Eastern and ancient Neolithic people uniformly share less drift with Yamnaya than Corded Ware does.

Ramber said...

Thank you again Davidski, but wouldn't you say the eastern Mongolians might have a similar genetic composition to Mongola? I mean aren't Mongola and Mongolians both Mongol groups?

Davidski said...

Ramber, I don't know much about Mongolia or the Mongola. But you're right in your observation that the Mongolian and Mongola samples we have are very different.

Alberto said...

@Balaji

Thanks. So I was a bit confused about this terminology. I thought that ANI was something intermediate between ASI and Modern North Indian (Pathans being the most Modern North Indian). So I supposed ANI being related maybe to Dravidian speakers or so (while ASI is quite different, more related to Onge and SE Asians).

Grey said...

Gioiello

"These words are from Latin, or better from some Neolatin languages, very likely Spanish, and don't demonstrate anything."

I'd say native words for gold and silver - then IE loanwords for copper, tin and bronze - and then a native word for iron demonstrates quite a lot.

Davidski said...

This is very interesting. There's a clear difference between the two Battle-Axe or Swedish Corded Ware samples. RISE98, which belongs to R1b, and has a somewhat dubious Battle-Axe classification, basically looks like a modern Swede, while RISE94, the confirmed Battle-Axe sample that belongs to R1a, clusters where Corded Ware samples are usually found on such PCA.

https://drive.google.com/file/d/0B9o3EYTdM8lQbTNGYXVFcVVsV00/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQUTlRTGxtd245UUE/view?usp=sharing

Gioiello said...

@ Grey
With due respect... but our analyses have to be careful, otherwise conclusions are wrong. I wrote to you also on Dienekes'AB about American languages. Linguistics is a hard science, very difficult, and there are the most weird ideas.
Spanish words entered Basque language in Medieval times.
The word for "iron" (burdina) is a word entered very likely around 1000 BCE, by judging from Latin "ferrum" > *fersom > *bherdom and Hebrew BRZL (barzel), i.e. is a wanderwort, i.e. a word brought to Basques from Phoenicians or other seafarers. The other words haven't a sure etymology... thus I think that your hypotheses, unfortunately, aren't supported by any proof.

Alberto said...

@Grey, Gioello

I think Gioello confused the Basque word for "lead" (berum) with the one for "iron" (burdina). I doubt that "burdina" is related to the Latin "ferrum" in any way, and indeed looks native.

More dubious are the words for gold and silver. Gold (urre) could come from the Latin "aurum" (Spanish "oro"), while silver (zilar) could be related to the English/Germanic one. But then again, if an expert says they're not related who am I to doubt his opinion?

Alberto said...

@Gioello

Sorry, I thought you were relating "berum" to "ferrum", but you were referring to the correct word "burdina". Your etymology about it could be right then, I didn't know the Phoenician connection.

rozenfag said...

Aleksandr Kozintsev posted this in Russian: http://генофонд.рф/?page_id=3877&cpage=1#comment-1066

My translation:
" Aleksandr Kozintsev:
21.06.2015 at 21:14

My anthropological commentary:

1) I have 21 Yamnaya craniological series in my database. No one of them is closer to South Caucasus than to local Eneolithic(Chalcolithic) groups - Sredny Stog and Khvalynsk. And no one is closer to Central Europe than to Sredny Stog and Khvalynsk. Therefore we have all the reasons - both morphological and genetic - to consider Yamnaya as autochtonous in steppe. The main intrigue is of course Eneolithic steppe. There are no genetic data, craniologically Sredny Stog and Khvalynsk look very distinct and there are no clear connections with either South or West. Typologists of course find Mediterranean influence and else, but this is illusionary.

2) Early Catacomb(Mediterrenean), as I wrote before in "indoiranian" discussion, seems to be migrants - either from Central Europe or South Caucasus, with equal probability. However Late Catacomb - clearly descendents of Yamnaya. How Early Catacomb wedge themselves between Yamnaya and Late Catacomb(in other words where did this migrants go after) - I have no idea.

3) Genetic data tells us that Central European Corded Ware related to Sintashta Bypassing Steppe. I can provide craniometrical parallel to that: West Kazakhstan Andronovo(Alakul-Kozhumberdy) have surprizingly European appearance - many early(Neolithic and Eneolithic) Central European parallels. From our series these humans related not to Srubnaya(as Alekseyev claimed) and not to Central Asians(as Ginzburg thought), but to some Yamnaya(for example from Ingulets river) and Early Catacomb(for example from Molochnaya river).

4) We are very happy that Allentoft et al. proved our conclusion about "Americanoidness" of Minusinsk Okunevo. Back then, in 1995, very few believed us, we waited for confirmation 20 years and finally we have it.

5) However Tuva Okunevo(Chaakhol) is completely different. They represent easternmost part of that migration vector, from the Steppe(Yamnaya and Catacomb) to the East, into Inner Asia. Quite possibly that they are one of Indo-Iraniian branches."

Grey said...

Gioiello

Leaving aside the argument over the word for iron, Basque having IE words for copper, tin and bronze is pretty interesting.

In the context of an idea that copper workers may have spread in advance of any steppic migration it doesn't prove anything but it certainly doesn't help to disprove it.

.

"Spanish words entered Basque language in Medieval times."

If you say so. I'd say that if this is true

http://www.inrap.fr/preventive-archaeology/Archeopages/Issues/SE01-21-22-23-2008/Issue-22/Dossier-Mines-and-quarries/p-11068-lg0-The-origins-of-mining-in-the-Western-Pyrenees-Digging-for-copper-iron-gold-and-silver.htm

"The first peak points to metal production in the first half of the 3rd millennium. Further to the east, in the Aspe Valley, a second peak relating to copper mining can be dated back 5,000 years and constitutes the oldest archaeological evidence available to us in the Pyrenees."

then the words for copper, tin and bronze might have carried over considerably earlier.

.

"thus I think that your hypotheses, unfortunately, aren't supported by any proof"

Well the word etymology part is from a Mr Trask.

http://www.buber.net/Basque/Euskara/Larry/metal.names.html

Alberto said...

Till we get better analyses from David, I was looking again at the D stats from the Allentoft et al. paper. In particular at the relationship between BA Armenian and Yamnaya/Afanasievo. Something interesting:

Yoruba baArm baAfan baYam -0.0071516804 0.0027932854 -2.5603114295 364707

The BA Armenian looks significantly closer to Afanasievo than to Yamnaya. We don't know yet how these new Yamnaya samples compare to the ones from Haak et al., but at least with this ones from Allentoft the direct link they establish from Yamnaya to Afanasievo doesn't seem to have strong support. It would be strange that in a migration to the east they would pick up more "Armenian" DNA.

Did this "Armenian" DNA enter from Central Asia to Afanasievo directly (vs. coming from the Caucasus)? Or something else could have happened?

Interestingly, CW is closer to Afanasievo then to Yamnaya, maybe meaning that the population that contributed to CW had more "Armenian" DNA?

Yoruba baCw baAfan baYam -0.0069193453 0.0026427011 -2.6182852949 506941

Also interestingly, in Unetice the Armenian affinity dilutes with respect to CW:

Yoruba baArm baCw baUne -0.0048692906 0.0025569935 -1.9043030999 310914

But in Bell Beakers it increases (very slightly, but still surprising):

Yoruba baArm baCw baBb 0.0003194889 0.002874605 0.1111418327 281024

Because mixing with more Neolithic European populations would dilute the affinity, it seems:

Yoruba baArm baBb baHu -0.009990499 0.0026235232 -3.8080467677 379834

This might relate to what Colin Welling has been suggesting, that West Europeans got their Yamnaya ancestry from a different source and not directly from CW? It had to be a Yamnaya that had quite a bit more "Armenian" than the one from CW, because BB have also quite more Neolithic that would dilute the "Armenian" affinity compared to CW, but instead it increases.

Gioiello said...

@ Alberto
[I think Gio[i]ello confused the Basque word for "lead" (berum) with the one for "iron" (burdina). I doubt that "burdina" is related to the Latin "ferrum" in any way, and indeed looks native] Sorry, I thought you were relating "berum" to "ferrum", but you were referring to the correct word "burdina". Your etymology about it could be right then, I didn't know the Phoenician connection.
More dubious are the words for gold and silver. Gold (urre) could come from the Latin "aurum" (Spanish "oro"), while silver (zilar) could be related to the English/Germanic one. But then again, if an expert says they're not related who am I to doubt his opinion?
------------------------------------------
Of course I didn't confuse anything, as you later realized. Anyway I did some simple observations with what I knew about the argument and unfortunately this isn't a blog of linguistics, anyway also linguistics may be a tool for understanding genetics and I used from the beginning of my theories. The Phoenician link is only an hypothesis, also for the expansion in -n of the Basque close to the Hebrew expansion in -l, but only a work hypothesis. That berum ("lead") could be linked with Latin *bherdom (and with its source) would be interesting to investigate.

Grey said...

Petter Johansson

"I think it is hard to make a model of all Uralic people spreading a Siberian component, since Mordovian people at the very Urheimat have so little Siberian and Mari have gotten it mainly from Turkic people."

Does it have to be all Uralic people? If it came about as a result of a particular group sitting on the Kargaly copper field then it might have been spread only by that specific population?

Assuming that original population (if it existed and still exists) was pushed off the steppe at some later point then is there a specific Uralic population living in the nearest refuge region to Kargaly (Urals or south?) with a high enough amount of this Siberian admixture?

Gioiello said...

@ Grey

Grey, I think that the true explanation is another and more simple: we know from history that Romans didn't rule Basques if not Basques romanized and that the Latinization of the whole Iberia happened very likely after the fall of the Roman Empire, thus Basques adopted those words from Spanish speakers in medieval times.

Mike Thomas said...

Balaji; Alberto

You;re right. Archaeologically at least, the Majkop-Yamnaya horizon appears to 'collapse' c. 2500 BC, segments, and retreats to the North Pontic region. In fact, the Catacomb culture - at least according to mtDNA, and Rozenfags kind translation - seems to have had rather large new impetus from elsewhere (quite possible some kind of Balto-Siberian forest people).

in central Europe and SEE, new influences appear in the mid-late Bronze age (ie c. 2500-2200 BC), ultimately from Near East. Perhaps correlated with J2 ? Correlated with inception of the Myceneans

Whatever, the case, if R1b-ANE is truly of steppic origin, it seems Yamnaya legacy nevertheless lasted genetically. But there are still some gaps in data, and ? apparent inconsistencies.




Grey said...

@Gioello

Well I certainly can't argue linguistics but if the French paper is correct

http://www.inrap.fr/preventive-archaeology/Archeopages/Issues/SE01-21-22-23-2008/Issue-22/Dossier-Mines-and-quarries/p-11068-lg0-The-origins-of-mining-in-the-Western-Pyrenees-Digging-for-copper-iron-gold-and-silver.htm

then copper mining in the Pyrenees goes back 5000 years.

Petter Johansson said...

"You can argue that Finns behave in this way because they have Saami admixture, and Saamis have Arctic admixture, while the Mordovians have Turkic admixture, and so on and so forth. But I won't be convinced, because I can see some very strong correlations there between language and genes."

I still think the actual spread must be modelled somehow for the correlation to be true. And, if Uralic peoples did indeed acquire Siberianness from different sources, it cannot be modelled as an ancestral component, or at least not very easily. A major problem is also that Arctic Uralic people in Europe have more Siberian admixture than Mordovians at the Urheimat, and Mordovians are hardly assimilated Slavs (on the contrary, Slavs assimilated Uralics). The spread of Uralic languages can maybe be modelled as several waves, where the most Northern one picked up and spread Siberian genes, if one really is forced to assume such a thing. Most likely IMO is simply that Siberian genes have been common in North-Eastern Europe for a long time, Uralic people not being an exception.

Grey said...

Wiki says (no idea if correct) the word "copper" comes from "Cyprus" which is interesting.

https://en.wiktionary.org/wiki/copper

Gioiello said...

@ Grey


I thank you for the link. Now beard and something for dinner, after I'll read the paper in French and I'll let you know. A link with Oetzi, hg. G, I2a-M26, Caucasian languages etc. could be interesting.
All the words for "copper" derives from "Cyprus", but Latin "aes" from IE *ayes etc.



Davidski said...

Alberto, here's what I'm getting. But my dataset is slightly different than theirs.

Yoruba RISE_baArm RISE_baAfan RISE_baYam -0.0048 -1.433
Yoruba RISE_baArm RISE_baAfan Yamnaya 0.0025 0.69
Yoruba RISE_baAfan RISE_baYam RISE_baArm -0.0404 -12.763
Yoruba RISE_baAfan Yamnaya RISE_baArm -0.0441 -11.741
Yoruba RISE_baAfan RISE_baCw RISE_baArm -0.0227 -5.631
Yoruba RISE_baAfan Corded_Ware_LN RISE_baArm -0.0322 -7.954
Yoruba RISE_baAfan RISE_baSin RISE_baArm -0.0146 -3.915

Yamnaya is Samara Yamnaya. So from that it looks like Afanasievo was derived from the Samara region, and Neolithic farmer ancestry increased and/or ANE dropped in the Corded Ware and Sintashta relative to Yamnaya and Afanasievo.

Nirjhar007 said...

David,
//I can't imagine R1a-M417 in Turkestan 5000-4500 years ago. Maybe, just maybe, it was beginning to move out from Europe into Asia along the steppe at that time.//
Imagine:).

Davidski said...

So it wasn't near the Caspian until Sintashta, but it was in Turkestan 5,000 years ago?

Why?

Alberto said...

@David

That's interesting. So it seems that Samara Yamnaya and Afanasievo had a similar amount of "Armenian", but the new samples from a later and more southwestern area of Yamnaya have less "Armenian".

That indeed makes more likely a migration from Samara to Afanasievo.

But why would the new Yamnaya samples, much closer to the Caucasus have more EHG?

Nirjhar007 said...

Are You delusional? and do we have any aDNA from those Parts from Copper Age-Bronze age to say it wasn't there?

Nirjhar007 said...

David,
BTW if the Sintashta R1a is indeed Z2124 then its clearly a SC Asian derived Y-DNA...

«Oldest ‹Older   1 – 200 of 244   Newer› Newest»