search this blog

Thursday, December 17, 2015

At least three genetically distinct Indo-European migrations into South Asia


First came the Indo-Aryans, probably in a couple of waves. Historical linguistics and archeology tell us that they originated on the Trans-Urals steppe in the Sintashta-Andronovo horizon, and pushed south around 2,000 BC to establish themselves as the ruling elite over Central Asian agriculturalists, who were probably in large part of West Asian origin.

There are multiple lines of genetic evidence suggesting that this is indeed what happened, which I discussed in detail in several earlier blog posts, like here.

But arguably the easiest way to show it is with D-stats of the form D(Indo-Aryan,Southeast_Asian; X,Outgroup), where the Indo-Aryans are the Kalash, a population isolate from the Hindu Kush with a relatively low level of extra-West Eurasian admixture and speaking an archaic form of Indo-Aryan. The Southeast Asians are the Dai from southern China, one of the best proxies for the South and East Asian admixture in the Kalash, while X represents a wide variety of present-day and ancient populations in my dataset. The top five D-stats, each based on well over 500K SNPs, are listed below:

Kalash Dai Kotias Ju_hoan_North 0.0684 22.704
Kalash Dai Sintashta Ju_hoan_North 0.0632 25.036
Kalash Dai Georgian Ju_hoan_North 0.0625 30.991
Kalash Dai Afanasievo Ju_hoan_North 0.0612 24.496
Kalash Dai Yamnaya_Samara Ju_hoan_North 0.0611 27.97

Really cool results. Obviously, Kotias is the recently published Caucasus hunter-gatherer (CHG) genome. The Kalash appear to carry the highest level of Kotias-related ancestry among present-day populations, which they probably acquired from both the Central Asian agriculturists and Indo-Aryan invaders. At the same time, however, Georgians show the highest affinity to Kotias because they harbor less extra-West Eurasian admixture.

After the Indo-Aryans came the Iranians, in all likelihood also from the steppe. They were either an offshoot of Sintashta-Andronovo or the more westerly Srubnaya Culture. I'd say the D-stats below, of the form D(Eastern_Iranian,Southeast_Asian)(X,Outgroup), are inconclusive, because the differences are small, and the outcome possibly affected by the methodology and/or sampling bias.

Tajik_Shugnan Dai Sintashta Ju_hoan_North 0.0716 26.427
Tajik_Shugnan Dai Poltavka Ju_hoan_North 0.0695 25.234
Tajik_Shugnan Dai Afanasievo Ju_hoan_North 0.0691 24.703
Tajik_Shugnan Dai Srubnaya Ju_hoan_North 0.069 28.266
Tajik_Shugnan Dai Corded_Ware_Germany Ju_hoan_North 0.0684 27.328

But again, the top five results make a lot of sense in the context of historical linguistics and archeology. By the way, Tajik Shugnans are a population isolate in the Pamir Mountains, like the Kalash with low level extra-West Eurasian admixture, and thus likely to be among the best available reference groups for early Eastern Iranians.

Interestingly, based on that list the Shugnans look more European than the Kalash. In large part this might be a reflection of the sharp rise in the level of European-specific Western hunter-gatherer (WHG) admixture on the steppe during the Middle Bronze Age, probably caused by population movements originating at the western edge of the steppe and/or in East Central Europe.

As far as I can tell, the fact that the Shugnans and Kalash have around the same level of extra-West Eurasian admixture means that I can try to hone in on the differences between their steppe-derived ancestry with D-stats of the form D(Kalash,Tajik_Shugnan)(Kotias,X). The top result seems to confirm my hunch, because Loschbour is, of course, a Western hunter-gatherer.

Loschbour 0.0149 3.874
Basque_Spanish 0.0113 4.232
Anatolia_Neolithic 0.0112 4.257
Karelia_HG 0.0105 3.005
Poltavka 0.01 3.539
Corded_Ware_Germany 0.0099 3.734
Afanasievo 0.0094 3.213
Srubnaya 0.0094 3.538
Yamnaya_Kalmykia 0.0091 3.362
Albanian 0.0088 3.419
Altai_IA 0.0088 3.087
Sintashta 0.0088 3.146
Greek 0.0076 3.094

Full output available here

More recently, during historic times, large parts of northern South Asia were settled by the Balochi, a Western Iranian people from the South Caspian region, whose ancestors were probably Indo-Europeanized a couple millennia earlier by Proto-Iranians from the steppe moving west across the Iranian Plateau. D-stats comparing the Balochi to the Kalash and Shugnans, respectively, clearly reflect the Near Eastern origins of the Balochi.

BedouinB 0.0104 6.151
Anatolia_Neolithic 0.0094 5.495
Druze 0.0084 5.228
Cypriot 0.0082 4.839
Syrian 0.0079 4.714
Armenian 0.0063 3.935
Satsurblia 0.0059 2.472
Georgian 0.0055 3.443
Iranian 0.0055 3.345
Abkhasian 0.0053 3.279
Greek 0.0052 3.166

...

Okunevo -0.0081 -3.552
Karelia_HG -0.0104 -4.666

Full output available here

Satsurblia 0.007 2.078
BedouinB 0.0051 2.277
...

Basque_Spanish -0.0073 -3.156
Mezhovskaya -0.0085 -3.045
Altai_IA -0.0092 -3.677
Scythian_IA -0.0092 -3.108
Yamnaya_Samara -0.0095 -4.092
Karitiana -0.0098 -3.501
Karasuk -0.0099 -4.322
Andronovo -0.01 -4.09
Sintashta -0.01 -3.951
Corded_Ware_Germany -0.0102 -4.34
Srubnaya -0.0106 -4.605
Yamnaya_Kalmykia -0.011 -4.511
MA1 -0.0118 -3.691
Okunevo -0.0122 -3.783
Poltavka -0.0125 -5.043
Afanasievo -0.0136 -5.235
Loschbour -0.0148 -4.201
Karelia_HG -0.0208 -6.537

Full output available here

In this analysis I used ancient samples from the recently published Jones et al. and Mathieson et al. studies, available on request from the authors and at the Reich lab website here, respectively. The present-day samples are from the Human Origins dataset, also available at the Reich lab website.

147 comments:

Alberto said...

Very interesting results. Probably the most surprising for me are the ones with the Kalash. Showing higher affinity to Sintashta than to Yamnaya/Afanasievo is very telling, and rather unexpected for me.

If the main difference between Yamnaya/Afanasievo and Sintashta is the Euro_MN admixture in Sintashta (basically Anatolia_Neolithic + WHG components, which have always looked pretty low in the Kalash, if present at all) while the components in Yamnaya/Afanasievo (EHG + CHG) have always been high in the Kalash, then this means that the Kalash must have some good amount of that Euro_MN admixture that's hiding in Admixture.

Though I wonder if Corded Ware (from Europe itself) would confirm this... Probably it would appear below Yamnaya/Afanasievo. And I'm not sure what that could mean. What is the real difference between Corded Ware and Sintashta? In theory, Sintashta looks even more Euro_MN than early Corded Ware, so Corded Ware should be in between Yamnaya and Sintashta, not below both. Maybe Sintashta does have "something" that increases affinity to S-C Asians and that we don't know what it is exactly?

Rami said...

Indo Aryans and Avestan Iranians were pretty much Aryanized BMAC peoples. In other words West Asian type people with considerable Steppe admixture.

By Antiquity you have a slew of Iranian groups on the scene as well Kushans, who derive from Tocharians. Some are mixed up with other SC Asians others still retain their Steppe DNA. So this period is still murky. The lingua franca for lot of these groups is Bactrian a fore runner to Pashto.

I don't think modern day Baloch people are Kurdish migrants, they seem to represent something more archaic. Just like Azeri Turks are essentially Turkified Iranians, the same peoples in Balochistan/Sistan went through a similar process. Their Anatolian EEF is way too low. Unless ofcourse original Kurds lacked Anatolian EEF which would make no sense given that proto Kurds ie Medes have been living in Northern Iran since the Median Empire in the Iron Age.

Davidski said...

It looks like Sintashta has farmer ancestry from the Balkans without the extra WHG admixture that affects Potapovka, Srubnaya, and Corded Ware. Can't think of anything else that would make a difference.

Sintashta
Yamnaya_Samara 0.589
Anatolia_Neolithic 0.411
chisq 3.936 tail prob 0.414759

Sintashta(2)
Yamnaya_Samara 0.566
Starcevo_EN 0.434
chisq 3.939 tail prob 0.414306

I'd say Andronovo would be even better than Sintashta if not for the minor East Eurasian admixture it has, which gets in the way here because of the Dai as one of the references.

Davidski said...

Ryan,

I don't think there's anything mysterious about the Balochi.

The fact that they differ from the Kalash and Pamir Tajiks by having inflated Near Eastern affinity and they speak a Western Iranian language pretty much betrays their origins.

Alberto said...

Yes, Mathieson et al. also showed Sintashta as Anatolia_Neolithic + Yamnaya, without extra WHG.

But probably that's the reason. I think that one problem with those stats is that higher BEA in the population in the 3rd position increases affinity to Kalash (by decreasing affinity to Dai). So maybe Iraqi_Jew would get even higher score, or similar at least.

It is difficult to get absolute affinity to S-C Asian populations with D-stats because of their ENA and BEA admixtures.

Davidski said...

I ran a lot of populations, including Iraqi and Iranian Jews and Abkhasians. They were all out of the top five, and below Corded Ware.

It's not as simple as the Basal Eurasian ratio. Populations will only score high with the Kalash in this comparison if they have very high CHG ratios, like Kotias and Georgians. But if not that, then they need other stuff, like high EHG and very low to non-existent East Asian and ASI.

I think Sintashta's relatively basal Balkan farmer ancestry helps, but this might not be a coincidence. The Kalash might actually have some of this as a minor component.

Rami said...

How do you explain the Brahui and Makrani peoples then?
Balochistan peoples are the most Near Eastern shifted of South Asians but regardless they lack elevated EEF Northern Iranians, Kurds , Azeris and Anatolian Turks do. NO doubt the Kurd like population which brought Baloch
to Balochistan had a genetic impact , but its not like a whole population moved en masse. The people living there have been living there for thousands of years. This fits in with the spread of the Mergarh culture of Balochistan (proto IVC) 7-8 Kya and is contemporaneous with Neolithic Iran.

Ofcourse Kalash, Tajiks will differ. Even the Pashtuns living in Balochistan now would have much more steppe input as well as more EEF based of the results I saw.

Alberto said...

"It's not as simple as the Basal Eurasian ratio."

Yes, certainly not that simple. Real shared ancestry between Kalash and the 3rd population counts a lot, no doubt. But for the case of Sintashta having higher affinity than Afanasievo I think the extra basal Eurasian makes enough of a difference (because I think that absolute affinity should be higher to Afanasievo).

But good results anyway, if you ran many different populations that those are the top 5.

Davidski said...

Rami, I think it's pretty obvious that the arrival of the Baloch language in South Asia within historic times came with a large population movement.

I suspect that you're focusing too much on ADMIXTURE output, which can be useful, but often it's about as useful as reading tea leaves. You really need to show more love to formal stats.

Here's the top five D-stats for the Balochis, with lots of populations tested and markers used.

Balochi Dai Kotias Ju_hoan_North 0.0602 21.583 510071
Balochi Dai Georgian Ju_hoan_North 0.0591 31.513 594924
Balochi Dai Iraqi_Jew Ju_hoan_North 0.0588 30.231 594924
Balochi Dai Anatolia_Neolithic Ju_hoan_North 0.0584 29.061 592940
Balochi Dai Armenian Ju_hoan_North 0.0583 30.587 594924

They're obviously in large part the descendants of recent migrants from Iran. I'm not saying they don't have any ancestry from the Mergarh people, but that's likely to be a minor component of their genomes.

Brahuis are basically the same. And yeah, I know they're Dravidians, but I don't think they have much in common with their Dravidian ancestors anymore. They just look like more ASI-shifted Balochis who speak Dravidian.

Brahui Dai Kotias Ju_hoan_North 0.0589 20.669 510071
Brahui Dai Georgian Ju_hoan_North 0.0569 29.686 594924
Brahui Dai Iraqi_Jew Ju_hoan_North 0.0569 29.086 594924
Brahui Dai Armenian Ju_hoan_North 0.0563 29.31 594924
Brahui Dai Anatolia_Neolithic Ju_hoan_North 0.0561 28.042 592940

I have no idea who today would be a good proxy for the Mergarh people? Maybe someone like the Chamar, who are basically a two way and fairly even mix of CHG and ASI?

John Thomas said...

So, Sintashta originated deep in Europe - if they carried 'Balkan farmer ancestry'.

Davidski said...

Yep, probably near the Carpathians, like the far western part of the steppe maybe.

capra internetensis said...

Tripolye didn't just evaporate, and did mix with steppe people, but their descendants didn't spread east at first - seems the most likely source

Rob said...

Yes that part of the Europe: western steppe, near the Carpathians, with Balkan farmer mix? You know- that big culture called CT that I've been going on about for 15 years ;)

la seƱora bibiloni said...

If so, the options would be... either a) this "western steppe, near the Carpathians, with Balkan farmer mix" people (Cucuteni-Trypolye for short) went all the way to Sintashta, got indo-europeanized and came back to indo-europeanize Europe. Or b) that "western steppe, etc." people spoke some kind of PIE before moving to the east and also grew some PIE offshoots into Europe at the same time they were moving to the East. Confusing

Davidski said...

Cucuteni-Trypolye women certainly didn't go all the way to Sintashta. What might have happened was that proto-Sintashta formed from a mixture of Eastern hunter-gatherer R1a men and Trypolye women in Sredny Stog, and then expanded east across the steppe and forest steppe to form Sintashta-Andronovo and the early Indo-Aryans/Iranians.

By that time, Corded Ware and Yamnaya had already expanded west deep into Europe, taking other Indo-European languages with them.

Rob said...

Yes I'm not postulating any CT invasion of India, that's for sure

la seƱora bibiloni said...

My own little pet theory always was that Cuc-Tryp / Sredny Stog had something to do with PIE, so I can't be disappointed by these findings :)

Rami said...

@David your wrong.

Firstly modern day Dravidian speakers in South India are not the same as the ones who originally spoke it lol . It was brought with Neolithic Farmers from Southern Iran. No denying Balochis are quite Near Eastern shifted but how do you explain their lack of EEF then?

Unless ASI rich peoples existed in Southern Iran also but that is doubtful unless ancient genome results from the region indicate otherwise.

People can manipulate data to suit their narrative, I will just wait till genome data comes out from the region. Or Formal papers come out.

Davidski said...

I just told you, Balochis don't lack EEF-related ancestry, because they show strong affinity to western Anatolian Neolithic farmers.

I'm not saying anything new here, because they have shown EEF-related ancestry in ADMIXTURE runs in various papers, like in the optimal K=16 run from Haak et al. 2015 (orange cluster).

http://2.bp.blogspot.com/-j1CbyFRCP-c/VnOqo4QVSmI/AAAAAAAAD1g/0eZVuOVAOYQ/s1600/K%253D16.png

And if you think I'm manipulating the data, then download it and show everyone here how I'm manipulating it.

Coldmountains said...

@Davidski


Baluchs are genetically and phenotypically still quite close to Pashtuns but havs less steppe admixture. Their R1a is mainly R1a-L657 which is Indo-Aryan/East Iranic and absent among Kurds, which are linguistically the closest relatives of Baluchs. The same is true for Brahui. Baluchs represent in my opinion rather something archaic and Pre-Indo-Europeans/Pre-Pashtuns of South Afghanistan/Southwest Pakistan were probably similar to them. We have no ancient DNA from this region but I would not be surprised if Mehrgarth and Pre-Aryan Baluchistan was genetically also more Near Eastern shifted than other regions in South Central Asia. Kalash and Pamiri live far away from Baluchistan and much closer to the steppe so it is not surprising that they have much more steppe input than Baluchs. But anyways thanks for the article and really hope to see ancient DNA from this region.

Davidski said...

But if there was Middle Eastern (as opposed to just CHG) and Anatolian Neolithic influence in Mehrgarh, then surely we'd see it in Indian Dravidian speakers with relatively high levels of West Eurasian ancestry too. But they just look like a two-way CHG/ASI mixture.

Balochis look like Pathans with more Middle Eastern ancestry, which makes sense considering their language. So this orange stuff they have came with both Eastern and Western Iranians, and especially the latter, long after Mehrgarh had disappeared.

http://2.bp.blogspot.com/-j1CbyFRCP-c/VnOqo4QVSmI/AAAAAAAAD1g/0eZVuOVAOYQ/s1600/K%253D16.png

postneo said...

Its ridiculous: You cannot claim there were 3 digitally discreet migrations when there could have been a constant diffusion.

Claiming to know what language they spoke is even more ridiculous.

Yes zombie populations of Kalash and pamir isolates are good data points but they are not fully representative of all that could have happened in the subcontinent.

Davidski said...

I'm guessing you overlooked the title, which says...

At least three genetically distinct Indo-European migrations into South Asia

Seinundzeit said...

I think David is absolutely right, the Baloch are predominately West Asian. At Anthrogenica, Kurd posted this qpAdm model for the Baloch:

70% Northern Iraqi Kurd + 15% Pashtun + 15% GujaratiD

chisq=0.950

tail probability=0.917

It's an amazingly good model, and it makes complete linguistic and historical sense.

Balochi is quite closely related to Kurdish and other "northwestern Iranian" languages (which is stating the obvious, as it is a "northwestern Iranian" language). In addition, the Shahnama mentions "Baloch warriors" in Gilan, northwestern Iran. Most importantly, the Baloch themselves believe that they have roots in West Asia (in their legends, Syria is the Baloch homeland).

Also, Baloch R1a levels are rather low for a South Central Asian population, and they have many Y-DNA haplogroups specific to West Asia (but absent in South Central Asia). In addition, if I'm not mistaken (I need Parasar right now, he knows this stuff), Baloch R1a isn't mainly L657. Rather, their R1a is usually of the same kind seen in Pashtuns (anyway, I think one can find L657 in West Asia).

So it is no surprise that, based on d-stats and qpAdm, the Baloch are predominantly West Asian, with some South Central Asian and South Asian admixture.

Although this is a topic that is completely subjective, and deeply vague/murky, I think I should mention that Baloch people are physically quite distinct from Pashtuns. It is not hard to tell them apart, at all. The diasporic Pakistani Baloch that I know are scarcely distinguishable from the Iranians that I know. On the other hand, many Pashtuns can pass as West Asians, but something is always noticeably "off". There is a vague Eastern European (Ukrainians come to mind) or often Northern Caucasian (Chechen) "vibe" to almost all lighter Pashtuns, which you don't see among the Baloch. Also, the Pamiri "look" (as epitomized by speakers of the Pamiri languages, and by many Tajiks) and the Turkic Central Asian "look" (as epitomized by Uzbek, Turkmen, etc) are quite common among Pashtuns, but nonexistent among the Baloch. And extremely dark Pashtuns overlap rather well with Punjabis, while extremely dark Baloch still look quite distinct from northwestern South Asians.

Basically, in Pakistan and Afghanistan, isolated tribal Baloch people tend to look like this:

http://3.bp.blogspot.com/-88CzRgzeL8E/Ufi434Gg7FI/AAAAAAAAHlo/InJl_f2us8M/s1600/salman+rashid.jpg

http://www.instablogs.com/wp-content/uploads/2012/07/baloch20___warna20323_6iXui_30125.jpg

While the Pashtuns who live right alongside them tend to look like this (Pashtuns from Kandahar or Pakistani Balochistan):

https://farm6.staticflickr.com/5131/5499432437_cf1f59e3fb.jpg

http://2.bp.blogspot.com/-MVb6Lz14LEA/UhWvOoIHHQI/AAAAAAAAHrI/RNvV-UOaiNM/s1600/Shaista+Khan+Kakar.jpg

Again, it's all subjective, and genetic data is what really counts, but I'm just saying that both populations look distinct (even when they live together, and even when they live in similar social and economic conditions).

Dilawer (Eurasian DNA) said...

Having ancestry as a Kurd from both N Iraq and SE Iran, and being intimately familiar with Baloch, I could not help chiming in.

@Sein

Thanks for posting that. I had forgotton that I even ran that.

@David

You have correctly pointed out the Balochis roots from the area of NW Iran and the Caucuses. Some of the Balochis are relatively recent Kurd migrants (1500s) from NW Iran, and go by Balochi, while many have retained their Kurd title. Also in Pakistani Balochistan some Baloch use Kurd as their family name. An internet search for 'Kurd Balochistan" will produce some hits.

Some info on the Kurd - Baloch connection:

http://www.cpp.net.pk/2013/06/24/story-of-the-baloch-by-professor-naela-quadri-baloch/
https://en.wikipedia.org/wiki/Yarahmadzai_tribe

According to Balochi tradition and historians, many Balochi tribes were originally Kurds from the area of NW Iran and form Syria.

I ran some Dstats to show how simnilar Baloch and Kurd really are. These show Kurd more similar to Baloch than even the Brahui or Makrani are to them.

Most are used to Admixture results showing a shared Baloch/Brahui/Makrani cluster defined by the "Balochi" component. The reason for the disconnect between ADMIXTURE and these Dstats is simple. Clusters in ADMIXTURE are based on shared allele frequencies at various loci. These frequencies are shaped by recent genetic drift, and that is what ADMIXTURE picks up on.

Dstats on the other hand are a direct marker to marker comparison, with an output dependent on the number of matches at various loci, adjusted for drift. These are sorted with most similar to Balochi on top. Z scores were not posted as they were mostly at -100

POP1 TARGET POP2 OUTGROUP D SNP
Mbuti Kurd_Iraq Balochi Gorilla -0.374 85170
Mbuti Georgian Balochi Gorilla -0.3733 85301
Mbuti Kalash Balochi Gorilla -0.3729 85301
Mbuti Armenian Balochi Gorilla -0.3718 85301
Mbuti Brahui Balochi Gorilla -0.3718 85301
Mbuti Armenian Balochi Gorilla -0.3718 85301
Mbuti Pathan Balochi Gorilla -0.371 85301
Mbuti Greek Balochi Gorilla -0.3699 85301
Mbuti Tajik_Pomiri Balochi Gorilla -0.3695 85301
Mbuti Pashtun_Afghan Balochi Gorilla -0.3692 82372
Mbuti Turkish Balochi Gorilla -0.3687 85301
Mbuti Makrani Balochi Gorilla -0.3678 85301
Mbuti Iranian Balochi Gorilla -0.3676 85301
Mbuti KOTIAS Balochi Gorilla -0.3675 73801
Mbuti Druze Balochi Gorilla -0.3658 85301
Mbuti Burusho Balochi Gorilla -0.3656 85301
Mbuti Tajik_Afghan Balochi Gorilla -0.3628 82372
Mbuti Anatolia_Neolithic Balochi Gorilla -0.3626 85027
Mbuti Syrian Balochi Gorilla -0.3538 85301
Mbuti Puliyar Balochi Gorilla -0.3486 81714
Mbuti Paniyas Balochi Gorilla -0.348 41068
Mbuti Onge Balochi Gorilla -0.3381 41068
Mbuti .MOTA Balochi Gorilla -0.1016 81230


Dilawer (Eurasian DNA) said...

These are for Brahui, sorted with most similar on top. Kurd_Iraq is the most similar.

POP1 TARGET POP2 OUTGROUP D SNP
Mbuti Kurd_Iraq Brahui Gorilla -0.3744 85170
Mbuti Georgian Brahui Gorilla -0.3722 85301
Mbuti Kalash Brahui Gorilla -0.372 85301
Mbuti Armenian Brahui Gorilla -0.371 85301
Mbuti Pathan Brahui Gorilla -0.3697 85301
Mbuti Greek Brahui Gorilla -0.3691 85301
Mbuti Pashtun_Afghan Brahui Gorilla -0.368 82372
Mbuti Tajik_Pomiri Brahui Gorilla -0.368 85301
Mbuti Turkish Brahui Gorilla -0.3677 85301
Mbuti Makrani Brahui Gorilla -0.3669 85301
Mbuti Iranian Brahui Gorilla -0.3666 85301
Mbuti KOTIAS Brahui Gorilla -0.3666 73801
Mbuti Druze Brahui Gorilla -0.3648 85301
Mbuti Burusho Brahui Gorilla -0.3643 85301
Mbuti Anatolia_Neolithic Brahui Gorilla -0.3618 85027
Mbuti Tajik_Afghan Brahui Gorilla -0.361 82372
Mbuti Syrian Brahui Gorilla -0.3535 85301
Mbuti Puliyar Brahui Gorilla -0.3468 81714
Mbuti Paniyas Brahui Gorilla -0.3466 41068
Mbuti Onge Brahui Gorilla -0.3367 41068
Mbuti .MOTA Brahui Gorilla -0.1002 81230

Alberto said...

@Kurd

Not judging about the Kurdish-Balochi connection, just a technical question. Balochi and Brahui have almost identical relation to other populations (pretty much as expected). And they are neighbours and for all we know pretty much the same genetically. And yet, they both have higher affinity to (in the same order) Kurd_Iraq, Georgian, Kalash and Armenian than between themselves. Doesn't this look like a bug (or limitation) of D-stats to you?

Dilawer (Eurasian DNA) said...

@ Alberto

I agree, that if Dstat outputs were not drift adjusted, they would show higher similarity between Brahui and Baloch vs Baloch and Kurd_Iraq. However, if we neutralize recent drift, and go back in time, Baloch and Brahui would be expected to shareconsiderably less affinity to each other than presently. This is evident from the languages. Personally I don't speak Brahui, but unlike Balochi, Brahui has been characterized as having Dravidian origins. Also, Dstat outputs have been found to be accurate when analyzing other populations. So if the output is inaccurate it should not be limited to the aforementioned analysis

Davidski said...

And yet, they both have higher affinity to (in the same order) Kurd_Iraq, Georgian, Kalash and Armenian than between themselves. Doesn't this look like a bug (or limitation) of D-stats to you?

Using Mbuti as the outgroup means that the stats will be depressed for any population that has Sub-Saharan admixture, and some of the Balochis do have a couple per cent. Makranis have even more.

Dilawer (Eurasian DNA) said...

@ David

"Using Mbuti as the outgroup means that the stats will be depressed for any population that has Sub-Saharan admixture, and some of the Balochis do have a couple per cent. Makranis have even more."

I agree, however I checked Kurd_Iraq's SSA levels from the ADMIXTURE run from my Eurasia 10 run, and they were about 1%, vs. 1.5% for the Balochi average, and 3.5% for the Makrani average.

The 0.5% in SSA difference between Kurd_Iraq and Balochi is based on ADMIXTURE, so I figured I would go ahead and check affinity to Mbuti of Kurd_Iraq vs. Balochi/Brahui Makrani.

Surprisingly, Kurd_Iraq showed more of a tilt towards Mbuti than Brahui and Makrani. Here is what I got.

Kurd_Iraq Brahui Mbuti Gorilla 0.0049 1.137 85170
Kurd_Iraq Makrani Mbuti Gorilla 0.0047 1.117 85170

So we can rule out SSA as the causative factor for any bias in the stats I previously posted.

To further corroborate everything, I did a direct comparsion of Kurd_Iraq and Brahui vs Balochi. Here is what I got.

POP1 POP2 TARGET OUTGROUP D Z SNP
Brahui Kurd_Iraq Balochi Gorilla -0.0038 -0.944 85170
Brahui Kalash Balochi Gorilla -0.0026 -1.402 85301
Brahui Armenian Balochi Gorilla -0.0009 -0.543 85301


Brahui Pathan Balochi Gorilla 0.0005 0.385 85301
Brahui Greek Balochi Gorilla 0.0012 0.789 85301
Brahui Tajik_Pomiri Balochi Gorilla 0.0017 0.98 85301
Brahui Pashtun_Afghan Balochi Gorilla 0.002 1.174 82372
Brahui Turkish Balochi Gorilla 0.0034 2.593 85301
Brahui Iranian Balochi Gorilla 0.0052 2.954 85301
Brahui Makrani Balochi Gorilla 0.0061 4.73 85301
Brahui KOTIAS Balochi Gorilla 0.0072 1.7 73801
Brahui Balochi_Iranian Balochi Gorilla 0.0073 1.666 85011
Brahui Burusho Balochi Gorilla 0.0076 5.023 85301
Brahui Druze Balochi Gorilla 0.0084 5.602 85301
Brahui Anatolia_Neolithic Balochi Gorilla 0.0092 4.034 85027
Brahui Tajik_Afghan Balochi Gorilla 0.0105 5.08 82372
Brahui Syrian Balochi Gorilla 0.0255 12.687 85301
Brahui Puliyar Balochi Gorilla 0.03 11.144 81714
Brahui Paniyas Balochi Gorilla 0.0372 9.847 41068
Brahui Onge Balochi Gorilla 0.0511 11.61 41068
Brahui .MOTA Balochi Gorilla 0.2979 70.889 81230

This shows that Kurd_Iraq is slightly more similar to Balochi than Brahui is, albeit the stat is not very significant.

Alberto said...

@Kurd

"Also, Dstat outputs have been found to be accurate when analyzing other populations. So if the output is inaccurate it should not be limited to the aforementioned analysis "

This limitation of Dstats has been known for quite a while, but I raised it in this case because it's the perfect example to show that it's a wrong/buggy behaviour.

The language they speak and what you call "recent drift" (I think you really mean admixture) is not important in this case. Looking at these stats:

Brahui Kurd_Iraq Balochi Gorilla -0.0038 -0.944 85170
Brahui Kalash Balochi Gorilla -0.0026 -1.402 85301
Brahui Armenian Balochi Gorilla -0.0009 -0.543 85301

And the ones you first posted, I think it's safe to assume that if you change the position of Balochi and Brahui in them you'll get the same results (i.e, Brahui will also be closer to Kurd_Iraq, Kalash and Armenian than to Baloch, even if they speak a Dravidian language.

I don't know if the authors of the program are aware of this or not. It's never mentioned in the documentation as a "known limitation", so it might be a pure bug they're not aware of, and if it gets reported and they manage to fix the algorithm so it becomes smarter and behaves as expected (in this case it should show Balochi and Brahui as closer to each other than to anyone else), it would be a great boost for Dstats that would actually become useful to compare all kind of populations reliably, and not just a small subset of them, because if you mix populations with different kind of admixtures you'll get consistently wonky results. (Yes, small amount of SSA also triggers these problems, especially if one population on the other side of the stat has ENA).

So maybe someone should report it to see what the authors have to say about it. If they are well aware of it and say it's an inherent limitation of the method and can't be fixed, fine. But in this case they should also warn about it clearly and explain which populations you cannot mix or you will get wrong results.

Shaikorth said...

Kurd, can you check IBS sharing of Baloch, Brahui and Kurd_Iraq using those 85301 SNP's? If the result is the same as in d-stats the fault likely doesn't lie with the algorithm.

Krefter said...

I'm wondering, if CHG-EHG was sex-biased should the X-chromosome of Yamnaya be more CHG than the rest of their DNA? The same goes for modern European's X-chromsome, if Steppe-EEF/WHG admixture was sex-biased, European's X-chromsomes should be more EEF/WHG than autosomal DNA.

Balaji said...

Davidski,

Congratulations on choosing just the right kind of statistics to compare the Kalash, the Balochi and the Tajik_Shugnan. You are certainly correct that the Balochi has more West Asian ancestry that the Kalash and that the Tajik_Shugnan has more European ancestry than the Kalash.

However, the data that you have yourself presented do not support the notion that the people of Baluchistan descend predominantly from “Western Iranian people from the South Caspian region”. You had the following statistics.

Balochi Dai Kotias Ju_hoan_North 0.0602 21.583 510071
Balochi Dai Georgian Ju_hoan_North 0.0591 31.513 594924
Balochi Dai Iraqi_Jew Ju_hoan_North 0.0588 30.231 594924
Balochi Dai Anatolia_Neolithic Ju_hoan_North 0.0584 29.061 592940
Balochi Dai Armenian Ju_hoan_North 0.0583 30.587 594924

I doubt that Kurd_Iraq will follow the same pattern showing greatest affinity to Kotias – this seems to be a South Asian phenomenon. I think the extra Near Eastern ancestry of Balochi came with the Arab conquest of this region which happened in the seventh and eight century C.E. (The rest of the Subcontinent did not come under Islamic domination until the end of the twelfth century). This is supported by the following.

Kalash Balochi Kotias BedouinB 0.0104 6.151 510071

Kurd,

You have presented some interesting information showing the similarity of Balochi and Kurd_Iraq. Could you find the top five populations for D(Kurd_Iraq, Dai; X, Ju_hoan_North) or perhaps share Kurd_Iraq with Davidski?

Also Chad Rohlfsen found the following.

Anatolia_Neolithic1 Kotias Brahui Primate_Gorilla -0.0068 -1.990 22107 22409 454082

Could you calculate the following for comparison?

Anatolia_Neolithic1 Kotias Kurd_Iraq Primate_Gorilla

Tobus said...

@Alberto:I don't know if the authors of the program are aware of this or not. It's never mentioned in the documentation as a "known limitation", so it might be a pure bug they're not aware of

DStats are a very basic calculation, I don't think there can really be a "bug" like you suggest. I think the "issue" here is that these stats aren't significant, and any perceived preference is just chance... add another 100 SNPs and the sign may well flip. Here are the ones I could run with 500k+ SNPs:

Brahui Armenian Balochi Primate_Gorilla -0.0013 -1.058 26938 27008 532298
Brahui Kalash Balochi Primate_Gorilla -0.0005 -0.34 26969 26993 532298

Look at the ABBA/ABAB counts - we're talking about 70 SNPs defining the entire result for Armenians, and 24 for Kalash. These are both within the standard error for a zero result - if you toss a coin 10 times and get 6 heads, that doesn't mean the coin is biased.

For the record, here are the runs with Brahui/Balochi reversed, they fit your assumptions:
Balochi Kalash Brahui Primate_Gorilla 0.0019 1.369 26969 26868 532298
Balochi Armenian Brahui Primate_Gorilla 0.0013 1.066 26938 26869 532298





Krefter said...

Can you run Urnfield_I0099 and Hinxton genomes through CHG K8.

Tobus said...

For completeness:
Balochi Brahui Kalash Primate_Gorilla 0.0023 2.354 26993 26868 532298
Balochi Brahui Armenian Primate_Gorilla 0.0026 2.659 27008 26869 532298

Alberto said...

@Tobus

Thanks for the stats and comments. So what you're suggesting is that we're just dealing with populations (Kalash, Armenian, Kurd_Iraq, Georgian, Balochi, Brahui) that are basically identical to each other so that stats are insignificant and could fall in any direction by chance?

Maybe. That could be further tested by comparing how they relate to other populations. I think that would reveal differences (not huge, they do all share a large amount of ancestry) that would not go well with what we get from direct comparisons.

But in any case, why do Dstats behave differently to IBS? For example, a population that shares a large amount of ancestry with EEF but has a small amount of SSA admixture would appear high in a list of populations by IBS sharing with an EEF but would show low affinity to it in Dstats. This is basically the phenomenon I'm talking about, and it's clearly seen in many stats. Whether it's a bug, a limitation, a design choice or whatever I don't know. But it certainly give strange results when comparing populations with different kind of admixtures.

Alberto said...

Here for example is an IBS list of Stuttgart genome (from LBK_EN) provided by Davidski some time ago:

https://drive.google.com/file/d/0B2ZfdVZaNXDxekV2anZpaEc4c0k/view?usp=sharing

Populations like East_Sicilian, Cypriot, Maltese, Moroccan_Jew, Iraqi_Jew, etc... are high in the list. But if you compare by Dstats to others like English,, Irish, Scottish, Norwegian, Polish or Lithuanian (which are lower in the list) you might get quite contradictory results. Something like:

Primate_Gorilla Stuttgart Iraqi_Jew Lithuanian

I think won't quite match the expected result from IBS sharing of both.

Simon_W said...

Slightly off topic, but in my last comment here, in the thread on Sardinians and their CHG-related ancestry, I speculated that the Greeks may have originated somewhere near Armenia and that the Etruscans had roots in a pre-Greek Aegean population low on "teal". But recently I've noticed that the fineSTRUCTURE analysis in the latest Busby et al. strongly contradicts this theory. I'm referring to Figure S3. It's striking that the South Italian cluster itali8 and the North Italian and Tuscan clusters itali13, tsi23 and tsi70 have very different relations to the East Med area. The South Italian cluster has some sharing with Southeastern Europe, which is mostly composed of Bulgarians and Greeks. Whereas the Tuscan and North Italian clusters have no sharing with Southeastern Europe, but they do have quite clear sharing with the Armenia/Iran population group (35 Armenians, 15 Iranians, 5 Kurds) and the Cypriot population group (12 Cypriots, 1 Syrian, 1 Druze). These different patterns were never apparent in ADMIXTURE analyses and PCAs. But they show up with fineSTRUCTURE. And they can be hardly coincidental, because they are in line with the different pre-Roman ethnic structure of Italy: Greeks in the south, and Etruscans in Tuscany and northern Italy. The conclusion is: the strong CHG-Teal layer in Greece and the BA movements from West Asia into Greece must have been non-Greek, possibly Etruscan-related.

Kurti said...

@Steinundzeit

From my own observation Baloch look predominantly Kurdish with an strong Indian vibe on them. Many of the so called "Balochi" on the Internet are actually Makranis or Brahui. Brahui and Makrani are so similar to Balochis as they could be counted as one. But a distinc difference is that many Makranis have a significant Sub Saharan African admixture to them which is often enough good visible.

Brahui on the other hand have simply more of the Indian Vibe to them. There must be a reason why these three groups have different ethnic names.

I have been saying this all along but some people didn't get it in their head. IMO The Balochi are predominantly of Kurdish descend (since they say they came from Aleppo which was from my sources once predominantly Kurdish, nowadays only ~25%).

Brahui and Makrani imo are simply Balochis themselves, just the differentation is based on the fact that Makrani are those "who have mixed with Sub Saharan people" and Brahui are those who brought the Indian admixture to the Balochi by mixing with them. The reason for why I don't think Brahui represent an ancient Dravidian speaking population of Pakistan and Balochi are simply descend of them, is because the closest groups of Dravidian speakers to Brahui are those in NorthEAST and East India, not the Dravidian speaking groups of West or Northwest India.

And I have an explanation for that. I know that many Muslims from Bangladesh and Northeast India left India for Pakistan when Pakistan was born as a country. They most likely settled in the rural areas of Balochistan.

And I honestly believe before the Dravidian speakers from India came to Balochistan the "proto Balochi" even more similar to the Kurds than they already are.

Kurti said...

"I agree, however I checked Kurd_Iraq's SSA levels from the ADMIXTURE run from my Eurasia 10 run, and they were about 1%, vs. 1.5% for the Balochi average, and 3.5% for the Makrani average."

The problem is, it is hard to talk about "really average SSA" admixture in Makrani since the Makranis do not have uniformly similar SSA admixture. There are Makranis who are 1/4 to 2/4 SSA while the majority of Makranis who are almost identical to Balochi. This is why you have Makranis who look almost like simple SSAfricans and other who look just Balochi.

Simon_W said...

Busby et al. also inferred a Levantine & North African admixture event for Northern Italy, dated to c. 700 AD. A date which screams "Islamic expansion"! And I'm not thinking of Muslim raids in northern Italy, but rather of Christian families from the Levant and Northern Africa taking refuge in the safe north of Italy, which was partly even Byzantine at that time.

The recent paper on the population structure of Italy showed only negligible Levantine and North African admixture in northern Italy, but that was with ADMIXTURE. And there may be some subtle geographical substructure not covered by the paper. On GEDmatch I've found an Italian match, apparently a north Italian, who scores like a Corsican in the MDLP K13 Ultimate calculator, that is, quite EEF-like. With the two pop approximation he scores like 53.4% French_South + 46.6% Lebanese_Christian.

Shaikorth said...

Simon_W, the Italy paper also includes tests for admixture using ALDER (linkage disequilibrium-based three-population test) which includes dating. I wouldn't say the dating of either ROLLOFF or ALDER is fully reliable, but it may provide some evidence to the scenario you suggest.

The strongest (most negative Z value) events they inferred using a pretty extensive European/West Asian/North African set are listed and differ quite a bit by region. The abbreviations should be evident for the most part (Kay is Kayseri Turks and Chu is Chuvash).

Northern Italy
http://oi64.tinypic.com/t6p5ef.jpg

Central Italy
http://oi63.tinypic.com/111i2cz.jpg

Southern Italy
http://oi64.tinypic.com/jfccas.jpg

Grey said...

@Krefter

"x chromosome bias"

Unless there was selection in place after the event (and in the opposite direction) you'd think there would have to be.

Dilawer (Eurasian DNA) said...

@ Shaikorth

Point well noted. I will run some allele sharing and IBS for various regional groups, and will post within 12 hours

@ Balaji

The stats you posted are not the most relevant to what is being discussed. I'll post some stats that directly involve the Baloch and Kurds

I will also report later on the what if any the mechanism is for Dstats to neutralize recent drift.

Incidentally, for the qpAdm Sein posted, I recall that the fixed paths for Kurd Baloch were such that would lead you to believe they are a near clade.

Dilawer (Eurasian DNA) said...

Here is the IBS tables for some relevant populations, sorted by most similar on top:

NO POPULATION Brahui
1 KURD_SE 0.71967
2 Pathan 0.71951
3 KURD_Iraq 0.71864
4 Georgian 0.71852
5 Armenian 0.71849
6 Balochi 0.71833
7 KOTIAS 0.71816
8 Iranian 0.71812
9 Makrani 0.71807
10 Tajik_Pomiri 0.71799
11 Pashtun_Afghan 0.71756
12 Kalash 0.71675
13 .MOTA 0.67257


NO POPULATION Balochi
1 Brahui 0.71833
2 Pathan 0.71807
3 KURD_SE 0.71767
4 KOTIAS 0.71706
5 Georgian 0.71702
6 Armenian 0.71685
7 KURD_iraq 0.71681
8 Tajik_Pomiri 0.71654
9 Makrani 0.71654
10 Iranian 0.71649
11 Pashtun_Afghan 0.71591
12 Kalash 0.71515
13 .MOTA 0.67137


POPULATION Makrani
Brahui 0.71807
Armenian 0.71764
Georgian 0.71759
Pathan 0.71748
KURD_SE 0.71731
KURD_iraq 0.71711
Iranian 0.71695
KOTIAS 0.71680
Balochi 0.71654
Tajik_Pomiri 0.71634
Pashtun_Afghan 0.71579
Kalash 0.71483
.MOTA 0.67279


POPULATION Pashtun_Afghan
KURD_SE 0.7208
KOTIAS 0.7207
Pathan 0.7205
KURD_iraq 0.7196
Tajik_Pomiri 0.7192
Georgian 0.7187
Armenian 0.7186
Iranian 0.7184
Brahui 0.7176
Kalash 0.7170
Balochi 0.7159
Makrani 0.7158
.MOTA 0.6757

Shaikorth said...

Looks like Kalash D-stat similarity to Brahui (Mbuti Brahui Kalash Gorilla) is elevated compared to IBS - something we see with certain homozygous populations - but Kurds are indeed solidly among their closest matches.

Alberto said...

Thanks Kurd.

@Shaikorth

Maybe in this case it's related to Kalash being homozygous? I don't know, and also in this case we're comparing populations which share a lot of ancestry so it might be less clear. Though I thought that it was a good case to point it out because Balochi and Brahui are obviously (if only for geographical reasons) very similar, so the stats showing higher affinity from Baloch to Kalash than to Brahui looked like a good example of this strange behaviour.

However, there are many other examples with other kind of population:

Lithuanian Georgian Pathan Mbuti 0.0051 3.294

I haven't really seen and IBS list of Pathan with other populations, but I would be quite surprised if I saw Lithuanian significantly higher in it than Georgian. Or:

Lithuanian Syrian LBK_EN Mbuti 0.0327 13.854

While true that Lithuanian appears higher in IBS from different LBK_EN individuals than Syrian, it's by a very small margin, not the huge amount of difference that the stat suggests (though maybe in this case Mbuti might be helping to increase the values?).

We could look for many examples. Maybe a stat like this:

Primate_Gorilla Iraqi_Jew Lithuanian Palestinian

Could also show the strange pattern (I can't know for sure without testing) due to Palestinian having some SSA admixture and would show that Iraqi_Jew is closer to Lithuanian than to Palestinian? Which would again be a good example (if it happens) of the phenomenon I'm referring to. Probably those running stats regularly can comment more about it or have many other examples.

Simon_W said...

@ Shaikorth

Interesting, thanks! According to these analyses the North African signal in northern Italy seems to be a little older, Roman age rather.

Regarding my Corsican-like match, I've just learned that his mother is Sardinian, only his father is North Italian, so no wonder he comes out Corsican-like. Hence, no surprise population in northern Italy, but at least it shows how reliable the new MDLP K13 works.

Kurti said...

However this comment caught my attention

"More recently, during historic times, large parts of northern South Asia were settled by the Balochi, a Western Iranian people from the South Caspian region, whose ancestors were probably Indo-Europeanized a couple millennia earlier by Proto-Iranians from the steppe moving west across the Iranian Plateau."

It is simply wrong and absurd to talk about "Indo Europeanization" of Western Iranic groups because they differ significantly from the Steppe nomads. We could use the same argument to describe EVERY modern Indo European group, some more some less.

I doubt that the ancestors of the Western Iranic groups came straight out of the Steppes and mixed with the locals, just like I doubt that the ancestors of modern Italic, Celtic or Iberian groups came straight out of Yamna or a culture alike.

No one hear was "Indo Europinized. IF the ancestors of the Iranic groups really came ultimately and only from the Steppes, much more reasonable is to talk about a slow admixing process where the "original" Steppe ancestry got deluded in the interface of BMAC and Sintashta in Central Asia already creating imo the Proto Indo_Iranians which was ~70% Sintasthta like, becoming more deluded towards Southwest to ~55%, which created a new Iranic sub culture known as Yaz and is said to be the homeland of Avestan, just like Corded Ware was an Indo European culture of it's own and not simply "Indo Europinized".

From the Yaz culture than the Proto West Iranic groups would start their journey into the western parts of the Iranian Plateau, Mesopotamia, Anatloia and Levant. Those West Iranic groups would probably be up to ~80% YAZ like but only have ~30-40% Sintashta ancestry remaining.

This also fits with the historic accounts in West Iranic history. They see Avestan homeland as their origin and not the Steppes.

So for the West Iranic groups not Sintashta should be taken as the ultimate source population to determine their "Indo Europan level", but the Yaz culture imo. Would have been good to have some DNA from their. Hopefully in the future.

"Indo Europinazation" would mean a group straight out of the Steppes took a journey into the Iranian Plateau, without layover cultures inbetween, and forced their language on the majority of the local populations. That is far from the reality and what we know from historians and archeologists who describe how the Medes and Persian arrived on the Iranian Plateau as simple nomads searching for grassland for their domestic animals and horses.

https://www.youtube.com/watch?v=opbzxhgxii8

Kurti said...

But will be very hard to find burials of the Yaz culture simply cause

" So far, no burials related to the culture have been found, and this was taken as evidence of the Zoroastrian practice of exposure or so-called sky burial."

What we only can hope for are skeletal remains.

Alberto said...

@Kurti

Yes, skeletal remains have been found from those "sky burials", but whether it's possible to get any DNA from them is a different story.

This is a short trailer of a documentary about Gonur Tepe excavations by Victor Sarianidi:

https://youtu.be/0LMvFI8gPAU

Davidski said...

This is very interesting. The Kalash may well be a mixture of CHG, Sintashta and ASI, with their unusually high level of CHG skewing some of the standard analyses.

https://drive.google.com/file/d/0B9o3EYTdM8lQS2FoNnl1RkNtYTg/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQcFpUcXEyeENIdTg/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQZ2plSWZGTEtDcUU/view?usp=sharing

Alberto said...

Interesting, yes. These stats seem related to another peculiar ones that I observed in Allentoft et al. Supp. info about shared drift with Ust-Ishim:

https://docs.google.com/spreadsheets/d/1qI3UYIGoy74hrw_sTAA5ckiZeBZKO1VSNynOOe45pOQ/edit?usp=sharing

BaSin (Bronze Age Sintashta) looks quite out of place, down there next to Jewish and Near Eastern populations. I thought it was probably some artifact of the quality of the genomes, because all other stats seemed to put them among CW and Northern Europe.

But who knows. Maybe they are telling us something and Grigoriev was not that wrong about its Near Eastern origins?

http://csc.ac.ru/news/1998_2/2-11-1.pdf

Alberto said...

Leaving stats aside (which might show some slight strangeness with Sintashta samples for whatever reason), an interesting thought came to me.

When Lazaridis et al. was published, they speculated about some hypothetical Hunter-gatherer population from Eastern Europe that would have received input from a MA1-like population and be some king of mix between WHG and MA1. And in Haak et al. 2014 we get to see those samples, which means they already had the EHG samples by the time they published Laz. et al 2013 (but obviously not the Yamnaya ones yet).

Then when they published Haak et al. 2014, they mentioned an "Armenian-like" population possibly from the Caucasus as the admixing one. Then David Anthony revealed the results from the Khvalynsk genomes that we got almost a year later, and suggested that the "Armenian-like" population might be a Mesolithic one from the Caucasus. One year later, just shortly after we get the Khalynsk genomes, we also get the Mesolithic genomes from the Caucasus that are a good match for Yamnaya admixing population. So it looks like they had already tipped Mr. Anthony about preliminary results.

I recently checked that it's over a year since Nick Patterson said in an interview that they had already taken Maykop samples, but sequencing was pending. This probably means that by the time they published Mathieson et al, they had not only Maykop results for a while, but probably many others.

Why do I mention all this now? Because there was a rather strange and apparently quite speculative comment in that paper about the origin of Sintashta. I remember commenting about it and asking myself why they speculated about such possibility instead of going for the most obvious one, and if they knew something that we didn't.

In the end I forgot about it because it seemed like a slip. But thinking about it, how likely is that a paper read by over a dozen people several times before publishing would have such comment going unnoticed? And how likely is it that none of them knew that R1a-Z93 is a subclade of R1a-M417 present in CW? Not likely at all. And how likely is it that they were just speculating for the sake of it and not actually knowing perfectly what they were saying because of the yet unpublished results they have?

So I honestly don't know what they meant in that paragraph about Sintashta not coming from CW, but I'm pretty sure we'll find out the exact reasons in their next paper.

Davidski said...

Patterson didn't have any Maykop samples when he did that interview. He actually said so in the comments here. He, or rather Reich, got them from Oxoford Uni afterwards.

You might be right though that they know where Z93 came from. But it's probably from western Yamnaya, and I reckon the samples that they tested are these...

http://eurogenes.blogspot.com.au/2015/03/population-genetics-of-copper-and.html

FrankN said...

@Simon_W: As concerns Northern Italy, the impact of French/ Spanish rule in Southern Italy should not be overlooked. Due to temporary Arab control, there were huge Jewish and Muslim communities living especially in Sicily and Apulia. Norman rule had been very tolerant, a tradition partly continued under Staufen rule, to the point that Frederick II was excommunicated as being suspect of adhering to Muslim faith.
https://en.wikipedia.org/wiki/Muslim_settlement_of_Lucera

Subsequent French and Spanish rule brought inquisition, forced conversion, and expulsion / slaughtering of Muslim/ Jewish communities (e.g. Naples 1293, 1541). Prime emigration target, as before for Iberian Jews/ conversos, was the Osman Empire (Turkey, Maghreb), but a substantial portion also went to Tuscany and beyond (especially Lombardy and Venice, also the Upper Rhine) - Europe's richest and most urbanised region by that time (Flanders came close, though).
Genetic similarity of West Sicilians/ Cypriots/ Druze/ Ashkenazi is well known. Arabs and Jews played quite a role, but the pattern started much earlier (Phoenicians, East Med obsidian traders, neolithicisation by "island hopping", etc.).

@Kurd: Thx for the interesting stats! If Balochi really originate from Aleppo, they should have encountered quite a number of Armenian and Georgian merchants there, which could further explain some of the results. I'd also assume some Crusaders having survived in the area. Out of curiosity: Could you run tests against "Frankish" populations? Whatever you have available - the closer to Boullion, the better (e.g. Walloons, Luxembourg, Lorraine, Champagne).

When it comes to Balochi/Brahui vs. Tajik/Pathans, possibly also Kalash, the following might deserve consideration, also as concerns East Asian admix:
https://en.wikipedia.org/wiki/Yuezhi

Alberto said...

Yes, I was not linking the Maykop samples to the comment about Sitnashta's origin. It was just to further illustrate that they do have results from samples quite before we see them published. So any comment that they make as a speculation is usually because they already know.

In this case, their insistence in EEF migrating east to the steppe as the origin of the Sintashta and related cultures seems to be because they already have the samples that show it. And the only source of EEF east from Central Europe and onto the steppe is probably Cucuteni-Trypillian.

Kurti said...

@Alberto

Good observation. Actually they were talking about a proto Maykop like population in the Steppes as backbone of Yamna which was the target of Cucuteni-trypilian and Samara H&G influx. Thats How I remember them mentioning it.

Davidski said...

The same pattern will be revealed at the western and eastern edges of the European steppe; EHG R1 males taking nearby EEF and CHG females and then expanding out of the steppe in all directions as somewhat different but closely related mixed populations.

This will basically be the story of the early Indo-European expansions told next year with new ancient DNA.

FrankN said...

@Kurti: "I doubt that the ancestors of the Western Iranic groups came straight out of the Steppes".

First consider this:
http://mbe.oxfordjournals.org/content/30/4/824.full.pdf+html

By 2000 BC, pigs with Central European aDNA start to appear in the East Anatolian archeological record (Lidar Hƶyuk). By Hellenistic times, European pigs have almost completely replaced Anatolian domesticated pigs between Western Anatolia and Armenia, though not in Georgia and Iran, which continue exclusively breading Anatolian pigs (Fig. 1, p.827). Initially, the European pigs are accompanied by those with Anatolian Y2 haplogroup - rare in the EN/MN record (1 find each from Central Anatolia, LBK, Gumelnita), but frequent among North Pontic wild boar aDNA (c.f. link below).
A similar replacement is recorded in Romania. Here, Anatolian Y1, which, with some admixture of Balkan E1-A, dominates up to Cucuteni and Gumelnita, is around 2000 BC abruptly replaced forever by Central European E1-C (Wietenberg Culture, a Carpathian variant of Usatovo Culture).
http://rstb.royalsocietypublishing.org/content/370/1660/20130616

Secondly, a paper on admixture in Armenians that Dave linked here a couple of months ago shows, as the most recent admixture, "Northern Europeans" around 2000 BC. Strange coincidence. Kura-Araxes crumbled when? 2000 BC. First Akkadian report on Hithites? 19th century BC.

So, there were apparently pig farmers coming through the steppe, but ultimately originating north of it some time before 2000 BC. The closest proxy we are having for them now (which isn't neccessary the best one) is CW. Leaving Georgia unaffected (pigwise), they should have entered along the Caspian coast and apparently followed the existing trajectory of the Kura-Araxes culture.
Were they also ancestors of Western Iranians? I don't know. If so, they must have been accompanied by cattle breeders, and only the latter moved on to the Iranian Plateau.

For all interested in pig (a)DNA, here a few more recent studies on Iberia (77% based on local E1-A), pre-Hispanic Canary Islands (no Anatolian aDNA), and the current European distribution of E1-A (Iberia, Greece, Balkans) and E1-C (Central & Eastern Europe) among wild boar.

http://www.ncbi.nlm.nih.gov/pubmed/22221023
http://www.gsejournal.org/content/47/1/40
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0091401

The latter distribution is quite puzzling - I'd say E1-A pigs were much earlier domesticated than commonly assumed, and already shipped through the Mediterranean by the Epi-Paleolithic at latest. Note also that a genetic bottleneck during the Ice Age has been documented for Italian wild boar (E1/ E2), but not for E1-A and E1-C. The E1-C refugium is yet unknown - it obviously was neither the Mediterranean, nor the Balkans, nor around the Black Sea. Hungary? Western France?

Rob said...

Davidski

The mating patterns of steppe males seem very interesting
Do you plan on writing a paper on it ?

Davidski said...

I'm planning to do a paper on something. Not sure what yet though, because it depends on what's already covered in the papers that introduce new ancient DNA data.

It's very likely that the next paper that will come out from Harvard/Broad MIT focusing on the ancient steppe will cover this issue, because it's crucial to mapping out the origins and spread of PIE.

But in any case, I'll put up a post next year here asking for ideas for a paper and for people to get involved, if they have ideas and skills to offer.

Rob said...

I think isotopic data would be important also
But it's all a bit complex and often results are equivocal

Kurti said...

David said

"The same pattern will be revealed at the western and eastern edges of the European steppe; EHG R1 males taking nearby EEF and CHG females and then expanding out of the steppe in all directions as somewhat different but closely related mixed populations."

Doubt, the moment more samples appear you will simply see how wrong you are.

If that was the case we wouldn't see CHG specific yDNA in a rich Eneolithic Samara burial.

Kurti said...

I am sure if that crazy EHG male dominance theory was the case we would have heard of it already in the existing papers.

All I see is a single Haplogroup, r1b z2103 and later R1a z93 dominating the Steppe lands, no sign of any R1b L11 or R1a z283, the two dominating yDNA of Italo_Celtic, Germanic and Balto_Slavic groups.

Therefore it is in my view simply absurd to assume that the yDNA so far found in Yamna burials and later cultures is representative of the Indo Europeans as a whole.

It looks to me more like we are dealing here with two "lucky" family lineages who have established themselves as some kind of Elite and forced any other lineage out of the region very early.

A sign of this is simply the J Haplogroup and ~20% CHG admixture in Eneolithic EHG groups.

It just doesn't make sense in my eyes how the same people who allegedly didn't "accept" CHG males in the Yamna culture had CHG males among their lines in their original homeland around Samara.

You have J Haplogroup in CHG and J Haplogroup among EHG but it (and actually any other Haplogroup found to this day among EHG samples, irony is L23 or even z93 wasn't found among any EHG ) is missing in Yamna and the conclusion is that Yamna were all EHG males???

Is this meant to be a logical explanation? Certanly not in my eyes.

Davidski said...

There is no CHG-specific Y-HG in any rich Samara burials. The two Eneolithic Samara Khvalynsk males who were buried with copper belong to R1a and R1b and are almost pure EHG. The R1b guy does have an mtDNA that could be CHG though: H2a1.

And female mobility and exogamy among Bronze Age European cultures is well documented. It's really not very surprising to see it on the steppe, because it was common in Central Europe after it was overrun by steppe groups.

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

Everything fits.

Kurti said...

@FrankN

Interesting fact, the ancient Medes and Persians were always specifically talking about two regions as their homeland.

One is were Avesta is originating from. There are still debates where the book of Avesta really comes from, one is the Yaz culture (majority of scientists) another is somewhere around Urmia(allot of scientist take this also in consideration). I also tend to think the Yaz Culture is the origin of Avesta.

Another often mentioned region by Persians is what Cyrus calls "Umman Manda" and reffers to the region of Media which is located in between the Taurus mountains and Mesopotamia (basically modern "East Anatolia"). Cyrus speaks of Umman Manda as the original "Aryan" homeland. Take in mind Aryan was a term meaning "the noble, enlighted once", the West Iranic tribes tended to call the East Iranic (Central Asian nomads) Turans, meaning "people of the dark culture", not enligthened, still in tends living people.

Tur/Tar = dark
an suffix for plural

Tur/an = The dark(people)
Ari/an = the noble/englithened(people).


Of course the East Iranic tribes didn't like the usage of this term for them and preffered to call themselves Arians themselves.

But this was not the reason I mentioned it, the point is that when West Iranic tribes talked about "Aryans" they specifically meant the "West Iranic" group as mentioned above.

So when Cyrus said Umman Manda(East Anatolia) is the homeland of the Aryans, he might have meant the West Iranic groups who might indeed have formed out of Kura-Araxes culture.

As it looks today the whole Indo_Iranian branch seems to be not limited to one culture(Sintashta/Andronovo) as some people in the past might have thought.

Indo_Iranians seems to be the product of three/four geographically close and connected cutures.

Srubna culture representing the Cimmerian branch.

Sintashta/Andronovo East Iranic branch

BMAC the IndoAryan branch

YAZ culture East Iranic branches which fall outside the Scythian branch and in strong contact or stronger relationship to West Iranic groups.

Kura Araxes the West Iranic branch.

the Nuristani branch is still a mystery to me.


Now this would also explain why we have stronger levels and basal clades of R1b in West Iranic groups and in the region around the former YAZ culture.



Kurti said...

Dave said

"There is no CHG-specific Y-HG in any rich Samara burials. The two Eneolithic Samara Khvalynsk males who were buried with copper belong to R1a and R1b and are almost pure EHG. The R1b guy does have an mtDNA that could be CHG though: H2a1."

I think you did miss something(or did I?). There is a Eneolithic EHG burial in Samara who belongs to yDNA J and it is quite rich burial, speaking for an accepted member of the community

So once again, you don't let people in your house whom you don't even allow to set a foot in your garden.

Here you have signs of CHG in middle of your house(Eneolithic Samara) but refuse to accept that they also set a foot in your garden/land(Yamna).

It just isn't logical.

Davidski said...

There's no Y-HG J in any of the steppe remains from the Eneolithic to the Middle Bronze Age.

J shows up in a forager from Karelia and then in Kurgan burials near the Altai during the Iron Age.

This is what you and James are missing. You really need J to be found in at last one elite early Kurgan burial somewhere, or your theory about Proto-Indo-European being introduced to the steppe along with CHG isn't worth the bandwidth here.

Kurti said...

more about Umman Manda

"Umman Manda (Akkadian for host of Manda) is a term used in the early second and first millennia BC for a poorly known people in ancient near east whom by some scholars are identified as to be of Indo-European origin. They have been identified as in different contexts as Hurrians, Elamites, Medes, Cimmerians, and Scythians.[1] The homeland of Umman Manda seems to be somewhere from Central Anatolia to north or northeastern Babylonia in what later came to be known as Mitanni, Mannae and Media, respectively.[citation needed] Zaluti, a leader of Ummanda Manda is mentioned, whose name seems to have an Indo-Iranian etymology. He is even suggested to be identified with Salitis the founder of the Hyksos, the Fifteenth dynasty of Egypt.

The principal literary source is the so-called Cuthaean Legend of Naram-Sin, a composition that deals with the third-millennium king of Agade (Akkad) Naram-Sin and his struggles against the Umman-manda. As a literary topos, the Umman-manda represent a socio-cultural phenomenon with a strong theological basis: The Umman-manda are created by the gods and called forth from their homeland on the northeastern frontier of Mesopotamia by the chief god, be it Enlil, Marduk, or AŔŔur, for some particular work of destruction; since this destruction is divinely ordained, human beings are powerless to stop it, and in fact are enjoined against interfering; when the destruction is completed, the gods themselves will destroy the Umman-manda. In the literary topos, the Umman-manda is the enemy of civilization. The question of who the original Umman-manda were remains a mystery.[2]

In the first millennium BC, the term denoted Cimmerians[3] and/or Medes."


Central Anatolia(land of Hittites just a coincidence?) to Mesopotamia. And the mentioned names being obviously Indo_Iranian.

I also remember to have mentioned often that Herodotus as well Cyrus always mentioned the Gutians/Mitanni/Medes as the same people. Especially Herodotus mentiones the Medes to be the former Mitanni and also mentions a land called Matiene just slightly North of Media, being the homeland of the Cimmerians.

Kurti said...

@Dave

"J shows up in a forager from Karelia"

Than let it be a forager from Karelia but this damn individual was also autosomally predominantly an EHG, wasn't he? If J has reached all the way up to karelia how comes you assume there was no male contribution to Yamna just on the doorsteps?


And my other arguments still stay

"All I see is a single Haplogroup, r1b z2103 and later R1a z93 dominating the Steppe lands, no sign of any R1b L11 or R1a z283, the two dominating yDNA of Italo_Celtic, Germanic and Balto_Slavic groups.

Therefore it is in my view simply absurd to assume that the yDNA so far found in Yamna burials and later cultures is representative of the Indo Europeans as a whole.

It looks to me more like we are dealing here with two "lucky" family lineages who have established themselves as some kind of Elite and forced any other lineage out of the region very early.
"

Davidski said...

One stray Y-HG J in Mesolithic Europe way up north near Finland doesn't say anything about the steppe of the Eneolithic/Early Bronze Age or the Indo-Europeans.

There's no way of knowing what it represents. J might turn out to be a minor WHG lineage in Mesolithic Europe when more samples are tested. Or not. Makes no difference. There's no J in any of the elite Kurgans.

And I'm not sure why you're suggesting now that L11 and Z282 aren't from the steppe? Seems like a poor strategy considering that we have samples from Bronze Age Central Europe carrying L11 and Z282 with elevated steppe admixture that wasn't there before the Corded Ware expansions from the east.

It's not like there's any J among the Corded Ware, Bell Beakers, Unetice etc. It only shows up during the Late Bronze Age in a Hungarian sample that has less steppe admixture than modern North-Central Euros.

Alberto said...

But one thing that is still quite a mystery is the genesis of CW. When Haak et al. was published, the hypothesis of a large migration from Yamnaya people was plausible because it fitted the autosomal structure of CW, and just a few samples from a single region could not represent the whole Yamnaya horizon Y-DNA wise.

But after, we've been getting more and more samples from the steppe and they all turned up R1b. This makes a mass migration from Yamnaya to CW not plausible. Unless western Yamnaya does turn up to be R1a, but I'd say that it's either R1b or R1a-Z93 (if the speculative comment from Mathieson et al. was really founded in samples they already had).

So where did CW come from? Where did they get the CHG wives from? Any parsimonious hypothesis?

Davidski said...

L664, Z282 and Z93 are sister clades, which probably means they were close to each other before their big expansions.

So wherever Z93 Potapovka, Sintashta and Srubnaya came from, that's basically where Z282/L664 Corded Ware came from as well. Probably Sredny Stog and/or successor cultures.

That's why all of these groups can be modeled in a similar way; 50-70% Samara Yamnaya, and the rest EEF and often WHG. Note that in the Mathieson et al. paper Nordic LNBA is almost identical to Sintashta in terms of ancestry proportions, and has some Z282.

Rob said...

Dave

"all of these groups can be modeled in a similar way; 50-70% Samara Yamnaya, and the rest EEF and often WHG"

Yes can be modelled that way. The reality might be slightly but importantly different.

Davidski said...

They carry sister Y-chromosome clades and match each other and Yamanya and Afanasievo best in terms of shared alleles in high resolution tests. And that's despite the fact that Corded Ware mixed a little with farmers in Central Europe. They look like one big family to me.

Afanasievo
Yamnaya_Samara 0.3614
Yamnaya_Kalmykia 0.3601
Poltavka 0.3598
Srubnaya 0.3493
Corded_Ware_Germany 0.3463

Andronovo
Sintashta 0.3495
Yamnaya_Samara 0.348
Poltavka 0.3462
Srubnaya 0.3461
Afanasievo 0.3455

Khvalynsk
Yamnaya_Samara 0.357
Poltavka 0.3555
Afanasievo 0.3503
Yamnaya_Kalmykia 0.3492
Srubnaya 0.3444

Poltavka
Yamnaya_Samara 0.3619
Afanasievo 0.3557
Yamnaya_Kalmykia 0.3535
Corded_Ware_Germany 0.3469
Srubnaya 0.3465

Poltavka_outlier
Srubnaya 0.3461
Poltavka 0.3455
Sintashta 0.3431
Yamnaya_Samara 0.342
Latvian 0.3397

Potapovka
Yamnaya_Samara 0.3434
Poltavka 0.3432
Srubnaya 0.3429
Andronovo 0.3426
Corded_Ware_Germany 0.3424

Sintashta
Andronovo 0.3452
Srubnaya 0.3437
Yamnaya_Samara 0.3409
Poltavka 0.3401
Afanasievo 0.3395

Srubnaya
Yamnaya_Samara 0.3505
Poltavka 0.3472
Afanasievo 0.346
Corded_Ware_Germany 0.3446
Sintashta 0.3443

Yamnaya_Kalmykia
Yamnaya_Samara 0.3571
Afanasievo 0.3568
Poltavka 0.3543
Srubnaya 0.341
Corded_Ware_Germany 0.3405

Yamnaya_Samara
Poltavka 0.3582
Afanasievo 0.3536
Yamnaya_Kalmykia 0.3533
Srubnaya 0.3453
Corded_Ware_Germany 0.3446

Kurti said...

David said

"So wherever Z93 Potapovka, Sintashta and Srubnaya came from, that's basically where Z282/L664 Corded Ware came from as well. Probably Sredny Stog and/or successor cultures.

That's why all of these groups can be modeled in a similar way; 50-70% Samara Yamnaya, and the rest EEF and often WHG. Note that in the Mathieson et al. paper Nordic LNBA is almost identical to Sintashta in terms of ancestry proportions, and has some Z282."


Yes z282 and Z293 are sister clades point given, therefore it is interesting that we haven't found yet any of this in the Yamna and successor cultures.

z283 might be a sister clade but how and since when does this indicate that both evolved in exactly the same region?

Last time I checked yDNA "I" is the sister clade of "J" does that mean it evolved exactly where J also evolved? No it just means that a group of IJ individuals mutated into I and another group into J. This could have happened in the same region, but also in two entirely different regions. So your theory has a flaw here.

But let's say you are right and z283 and z93 evolved in "the same region".

But aren't you speculating about the presence of an Haplogroup without having any evidence?

So let me get that straight, you ASSUME that z283 should be found in the Steppes, because it is a brother clade of z93, despite it not being found in any of the burials,but have your issues in accepting that the same theory for Haplogroup J, despite it's sibling I being found in one Yamna burial and itself being found as far north as Mesolithic Karelia?

Double Standards?

And don't get me even started about R1b L11.

So either we are going to by what we have on hand, if you insist on "facts" on hand, which means we don't have any L11 and z283.

Or we are going to speculate about the existence of other Haplogroups despite no evidences.

But please no bouncing back and forth depending on what fits best for the own theory.

FrankN said...

@Alberto: So where did CW come from? Any parsimonious hypothesis?
More lines of thinking than already a hypothesis, but anyway:

The geographic origin of CW is yet unclear. Archeologically, the most probable scenarios are
a.) Lesser Poland, and
b.) SE Baltic (Lithuania and surrounding)

Recent research seems to favour the Baltic option, so let's concentrate on this, well bearing in mind that any new aDNA may change the picture.

A main point for me is that it isn't sufficient to identify areas of admixture (e.g. Yamnaya, Sredni Stog). The area needs to have the economic potential to sustain a large population, otherwise there is no way to explain the profound genetic change from Central European MN to CW.
CT, e.g., has been estimated at some 36-50 thousand people at its peaktime (frequent settlement relocation every 50-100 years makes ist look much more impressive on settlement maps than it actually was). This is some 1-1.5% of contemporary total European population. Even if all of CT woman married EHG/CHG men to form Sredny Stog, its hard to get to the numbers that can explain EHG/CHG shares in CW.

Unless, of course, we have been missing important things in NE Europe. I don't mean steppe cattle herding here. This would hardly get us to the WHG/CHG/EHG numbers required - check out current rural population densities in Mongolia, and you understand what I mean. Instead, I am thinking about

a.) underestimating the potential of aquatic foraging, Maglemose/Ertebolle/Pitted Ware (Alpine pile-dwelling) type, apparently exported to the SE Baltic coast by the 6th, and further south (Upper Volga, Nemunas, Pripyat?) from the early 4th millenium onwards; and

b.) the so far poorly researched case of early neolithic millet farming in Eastern Europe and the Steppe; IOW the possibility of EHG in fact being CHG-enriched Volga-Uralian Neolithic Farmers (VUF).

I will later post more details on both - now there are Christmas preparations to do.
On millet farming, you may already consult its current extent in the former Soviet Union (link 1), a review on the archeological record so far (link 2), and a study on its possible domestication history (link 3).
Note that proso millet (panicum miliaceum)is the most draught-tolerant cereal and well adopted to cold climate. It hasn't been part of the Anatolian package (first Mesopotamian evidence from the Iron age), but nevertheless shows up early in Transcaucasia, Moldova and LBK. It disappears in Rƶssen, but continues with Lenyel (evidence from Cracow-Mogala).

http://www.agroatlas.ru/en/content/cultural/Panicum_miliaceum_K/map/index.html
http://link.springer.com/article/10.1007/s00334-008-0187-1/fulltext.html
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3258423/

Davidski said...

@Kurti

So let me get that straight, you ASSUME that z283 should be found in the Steppes, because it is a brother clade of z93, despite it not being found in any of the burials,but have your issues in accepting that the same theory for Haplogroup J, despite it's sibling I being found in one Yamna burial and itself being found as far north as Mesolithic Karelia?

I and J are sister clades that expanded from the same place something like 25,000 years ago. Z282 and Z93 are sister clades that expanded from the same place around 5,000 years, because they didn't exist long before that.

And considering that Z282 is found in Corded Ware, which is a culture and population closely related to Andronovo, Khvalynsk, Srubnaya, Yamnaya etc., then what's the problem?

You can't wriggle your way out of this with some semantics, considering that we've seen several major papers recently showing that there were massive migrations from the European steppe to Europe and Asia which basically created the modern European gene pool.

What are we even debating here? Seems like you're in denial.

Ryan said...

Not to detract from the current discussion, but David - you said that Dravidians are a straight up mix of CHG + local hunter gatherers (ASI). If CT mixed with early Indo-Europeans, would it make sense to consider CT as likely something similar? IE CHG + WHG - explaining the excess of WHG in Yamnaya?

Also, have you taken another look at Burusho samples lately? I recall you saying they fit well as a mix of Georgians and Naxi. Is that still the case? Is there any more to learn there now that we have CHG samples?

Also, what do you (or others) think about CHG with relation to language families? Was CHG linguistically homogenous, and if so, would their language be Kartvelian, or North Caucasian, or both?

Ryan said...

Oh, and one other quick question - any idea why Z282 drops of well to the East of CWC's full range? Is this Bell Beakers encroaching on CWC?

Nirjhar007 said...

Hi guys,
Excuse me for being a bit OT. But i want to know the general consensus regarding this study by B Gowrys, which found possible Y-DNA G, I/J from Jagodno, Wroclaw in Poland. It is of course interesting that the site belongs to Corded Ware of 2800 BC period!.
http://www.sciencedirect.com/science/article/pii/S0305440313000459

Rob said...

Nirj

It was minimal tested SNP
They found it could be I, or J or E; but leant toward I because it's a 'native' European marker

So not conclusive

Nirjhar007 said...

Rob,
Thanks :) .
Ryan
Was CHG linguistically homogenous, and if so, would their language be Kartvelian, or North Caucasian, or both?
Seriously?!!.

Davidski said...

Ryan,

Burusho are basically Pathans with inflated affinity to Amerindians, East Central Asians and Siberians. I think it's likely that they acquired their language via a Bronze Age migration from somewhere near the Altai, like maybe by the Okunevo people? Or from another Siberian-related people living within the Andronovo horizon.

https://drive.google.com/file/d/0B9o3EYTdM8lQbGY1eFZvQ1paZ1U/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQLTVpd2FlOGVBb1k/view?usp=sharing

Nirjhar,

It's a long shot that the Y-HGs at that Polish Corded Ware site, which was far from anything elite, included a J. You're clutching at straws. You're a bit of a straw clutcher like Kurti.

Shaikorth said...

Treemix suggests the eastern influence of Burushos is similar to Kalash.

https://lh3.googleusercontent.com/ma0-ZIw4UR_2Hh5DN30JEDEpLTp3AUlpty6b2yam7Fg=w949-h631-no
https://drive.google.com/file/d/0B9o3EYTdM8lQellnZXY0Y1pFNjg/view?pref=2&pli=1

There is more of it (or they are less Indo-European and Dravidian) so it will not get absorbed into Kalash and South Indian-centric components when ADMIXTURE creates those but at lowest K's it looks similar there too. Okunevo at least was more like Amerindian and ANE.

Kurti said...

David said

"I and J are sister clades that expanded from the same place something like 25,000 years ago. Z282 and Z93 are sister clades that expanded from the same place around 5,000 years, because they didn't exist long before that."

The IJ example was brought up to make you clear that two Haplogroups being brother clades doesn't automatically imply they came to existence in the same region. You say that I and J evolved in the same region 25000 years ago, well there are a dozen of scientist who would disagree with you and say that I evolved somewhere else than J.

But than how does this disprove my point anyways? Even if both Haplogroups evolved in the same region doesn't this imply that where I is there could and should be some J too?

In fact many people are arguing that J is eastern Eprigravettian.


"
And considering that Z282 is found in Corded Ware, which is a culture and population closely related to Andronovo, Khvalynsk, Srubnaya, Yamnaya etc., then what's the problem?"

Corded Ware plays no significant role for our discussion here about Yamna and it's Haplogroups. Corded Ware did also have quite some other Haplogroups which have not been found in Yamna so far among them at least one G or J.

Dave here it has turned and almost sounds like I am arguing for the exclusion of Haplogroups such as z283 and L11. In fact I am actually arguing for their presence as seen in my posts above. I am trying to explain to you on the example of these two Haplogroups how, what we have found so far in Yamna burials can be misleading us for now.

But I see it doesn't help to discuss about this. You will remain on your "wive" theory and defend it like a national property. Until we get our hands on more samples.

Kurti said...

To all the people who get me wrong here. I am proposing for a long time now, that what we call "Proto Indo Europeans" are in fact a complex of related cultures. Not all the other being descend of one (Yamna).

In my own theory, Maykop, Kura-Araxes, cucuteni-tripolye, Sintashta CW and probably also Leyla Tepe are as much "PIE" cultures as Yamna.


Think about it for a second. Everything what we consider as signs of "PIE" culture have been found in all of the cultures above.

By 3300 BCE we find Horse bones in mount/Kurgan burials in Kura Araxes culture that predates Sintashta by 1200 years.

Kura-Araxes has also the other typical Indo European attributes. They were a mixed herding/farming culture and had wagons also.

https://en.wikipedia.org/wiki/Kura%E2%80%93Araxes_culture


Scientists are speculating a connection to Mitanni. Since Mitanni and Medes are equalized by too many ancient authors to ignore. In my opinion Mitanni/Medes and Persians might have their origin in Kura-Araxes. the Anatolia branch, if not Maykop than also Kura-Araxes. Many scientists are actually supporting a Maykop/Kura-Araxes origin of Hittites/Anatolian Indo Europeans.

Fact is Kura-Araxes, Maykop, Yamna and CT had at least close cultural and allot of genetic connections it seems.

Davidski said...

You don't have a clue what you're talking about.

PIE looks like a language family that developed in a relatively narrow area between the northern forests and the Caucasus.

It couldn't have been widely spoken until it broke up into several daughter branches. But by that time it wasn't PIE.

Moreover, even Renfrew was forced to put the first Indo-Iranians on the steppe, and that's saying something, because his Anatolian hypothesis is otherwise a crock of shit. And no one ever said Sintashta was PIE. It's Proto-Indo-Iranian.

Alberto said...

@FrankN

Leaving aside the CW origin debate here, but concentrating on the point you mention about population densities, I think that the main problem about population turnovers (which was my main concern about the steppe as a source of these turnovers) has been mostly solved now.

The biggest problem was Asia. But now it looks quite clear (pending samples to know with certainty) that Asia didn't experience any population turnover in the LN/BA period. If IE languages indeed entered Asia from the steppe, they did so without having a very significant genetic impact. So this problem is mostly gone now.

For Europe, we should distinguish southern and nothern Europe. For nothern Europe, a big genetic impact from the steppe (even more when steppe here means a mix of native steppe + CHG population) was always possible, so it wasn't a big problem in itself. Now with the latest numbers we could estimate it at about 40-45% in NE Europe and 30-35% in NW Europe. But even these estimates are pending a serious correction once we get pre-CW DNA from the CW areas. A good amount of the EHG admixture could come from local HGs, so that would reduce the steppe impact by some 10 points. So those figures would look reasonable for those areas (only the LBK area would look like it had a large impact for the kind of population it had, but a combination of factors, like local decay, and the movements being throughout all of the rest of northern Europe would make it sensible anyway).

For southern Europe, it doesn't look like the steppe was a source of any population turnover. The biggest change comes from CHGs, who came from Asia (which did have the numbers to have a big impact), so this is no longer a problem either.

More samples will help us to fine tune all the figures, origins, etc... Looking back, when the Lazaridis et al. model was introduced, it looked pretty horrible (I was quite critical with it back then). Then the Haak et al. model was a huge improvement with the addition of the "Armenian-like" population to the "steppe" and the MN European samples (with the Hunter-Gatherer "resurgence"). It still overestimated the so called steppe admixture in Europe (especially southern Europe), and while it didn't give figures for Asia, the blogosphere quickly did the math, bringing even worse numbers there. Now with the CHG genomes it's been another huge step forward and we're starting to get to reasonable numbers. So this means that the more samples we get, the more sense things make. Soon enough everything will start to fall into place quite nicely.

There are still many open questions and huge gaps in the data, but 2016 is going to be an amazing year for ancient DNA that will bring us a lot of answers and quite a few surprises.

Merry Christmas everyone!

Nirjhar007 said...

Merry Christmas to You guys!.
Dave,
You certainly think whatever sample comes from CWC will be R1a? or R1b?, i don't think its a practical approach mate.

Rob said...

Alberto

"For nothern Europe, a big genetic impact from the steppe (even more when steppe here means a mix of native steppe + CHG population) was always possible, so it wasn't a big problem in itself. Now with the latest numbers we could estimate it at about 40-45% in NE Europe and 30-35% in NW Europe. But even these estimates are pending a serious correction once we get pre-CW DNA from the CW areas. A good amount of the EHG admixture could come from local HGs, so that would reduce the steppe impact by some 10 points"

I suspect this mihjt be more than 10% when we get more samples from eastern Lengyel, GAC, T-C and some eastern "crypto-Neolithic" cultures

In fact, it might be along the lines of 35-40%

Merry Christmas y'all

Davidski said...

Estimates of steppe ancestry are too low for much of Europe currently and don't match the uniparental impact in many areas because the Yamnaya samples we have are unrealistic references. I suspect western and central steppe populations will look more like Srubnaya than Yamnaya Kalmykia/Samara.

In Asia they'll range from something like 70% in parts of the Pamirs to nothing deeper in India and the Near East.

FrankN said...

@Alberto:

I am not sure that we can explain CW simply from "some" Steppe admix.
This has most likely already been done somewhere, but I don't recall the figures: Could somebody try to display CW_Germany as a mix of CentralEur_MN, PittedWare, and Yamnaya, adding extra CHG or EHG, if required?
I suppose we will be getting close to 40% Yamnaya here. That would imply the migrating part of Yamnaya being as or more sizeable as CE_MN.

MĆ¼ller (link below, Fig. 17.9, p. 211) estimates the total population of Central Europe & Southern Scandinavia (starting from Moravia and the Alps, including Poland, but excluding Hungary) at some 2.2 million by 3500 BC. 700 thousand may be post-Ertebolle/ Swifterband (including Northern Polish FunnelBeaker), and thus more WHG (& EHG?)-enhanced. This would leave us with 1.5 million "typical" MN between Paris and Bydgosz/Warszawa, Lyon and Bratislava.
By 3200 BC, we have another 800,000 people living there, partly fuelled by population growth and colonisation of new areas in Southern Scandinavia, partly probably from immigration (which would place the start of the demographic change already into the GAC period). Trends differ strongly by region: Some earlier FB farming colonies such as the Gokhem area are given up, possibly for climatic reasons. Other areas, e.g. Eastern Sweden, become newly settled. Densely populated Danish Isles and Holstein keep population numbers stable by substantial outmigration, especially into Mecklenburg, Brandenburg, the Middle Elbe region, Pommerania and Kujawia.

https://www.academia.edu/10872287/8_Million_Neolithic_Europeans_Social_Demography_and_Social_Archaeology_on_the_Scope_of_Change_from_the_Near_East_to_Scandinavia

By 3000 BC, however, there are strong signs of reduced settlement intensity across Southern Scandinavia and parts of Central Europe. Reasons are not fully understood yet. Outmigration towards the Southeast, driven by soil exhaustion and colder climate may have played a role, but we may also have had preceding periods of famine and/or some epidemic. Emptied areas are soon filled up by CW- and PittesWare related immigration. This helps to contain the overall demographic drop to some 200,000 people, still 7% less than 2-3 centuries before.
Taking everything together, some 400,000 people immigrating between 3500 and 3200 BC (GAC), and maybe another 500,000 in the early 3rd millennium (CW/PW) look plausible.

Now, CW in Southern Germany/ Switzerland seems to primarily have been a cultural rather than a demographic phenomenon. The little archeological record I am aware of shows a CW village in Frankonia in good old EN/MN tradition - 6 houses clustered together on a hill, with farming land nearby and some grazing area in the valley below. Pile dwelling in Switzerland and Suebia continues as before. The only real change is appearance of stone-paved roads, wide enough for ox-carts, leading into the settlements. West of the Rhine, CW anyway wasn't much of an issue.
Hence, it appears that most of the demographic change didn't reach far beyond Elbe and Weser (or are we having aDNA telling otherwise?). This would mean the eastern half of CentralEur_MN, around 800,000 people by 3500 BC, encountering approximately the same number of EHG/CHG loaded immigrants over the next 600 years. So far, figures make sense.

But where did the 800-900 thousand immigrants come from? Clearly not CT. And I also have serious doubt that Sredni Stog/ Yamnaya could provide such numbers.

+++
Otherwise Happy Christmas to all of you (to the extent it has a meaning in your culture).

Davidski said...

I am not sure that we can explain CW simply from "some" Steppe admix.

Have you even bothered to look at the formal stats and uniparental marker data relevant to this question?

It won't go away just because you ignore it.

FrankN said...

@Dave: I dont think I am ignoring anything. Actually, a few months ago I had the opportunity to talk to the Custodian of the Hamburg Museum of Prehistory. He told me that he didn't need aDNA to figure out that there was a major population turnover around 3000 BC - a look at the skulls in his collection would do. And he is clearly not the only one. German publications on the MN - Chalcolithic transition have been full of alluding to "major demographic changes" around 3000 BC, well before this year's publications on "Steppe" aDNA.

However - to make it equally clear: There is nothing in the German archeological record suggesting anything like chariot-driving cattle herders conquering locals to bring about that change. On Eulau, which has provided a good chunk of the aDNA we are talking about, note that the CW settlers there have been massacred:
http://www.academia.edu/649220/The_Eulau_Eulogy_Bioarchaeological_Interpretation_of_Lethal_Violence_in_Corded_Ware_Multiple_Burials_from_Saxony-Anhalt_Germany

Prime suspect is the Schƶnfelder Culture, a late offspring from Norhtern Funnelbeaker -> Bernburg Culture, situated some 50 km north of Eulau. The Schƶnfelder Culture cremated their deceased, no chance for aDNA (unfortunately). Apparently, CW immmigration wasn't always welcome. The fact that Schƶnfelder Culture people massacred CW people, and not the other way around, however, is quite telling about the character of the processes that ocurred at that time.

Davidski said...

There is nothing in the German archeological record suggesting anything like chariot-driving cattle herders conquering locals to bring about that change.

Chariots were invented in the Sintashta Culture during the Middle Bronze Age, so Corded Ware people could not have driven them.

But apart from that, yes, Corded Ware cattle herders took over large swaths of Northern Europe during the Late Neolithic and brought about a massive genetic shift there. Ancient DNA shows this clearly.

On Eulau, which has provided a good chunk of the aDNA we are talking about, note that the CW settlers there have been massacred.

See, this is what I'm talking about. How is it that you haven't managed to move past 2009 yet, considering that you apparently read this blog?

You need to familiarize yourself with the latest data and make an effort to understand it, otherwise many of your comments here won't be relevant to the real world.

Alberto said...

No one is really denying the changes, it's more about understanding how they happened.

We can't just give for granted that the changes came from a massive migration from the steppe, for different reasons:

- The steppe samples we have are all R1b.
- While they match the (autosomal) genetic profile of the changes in Northern Europe, we lack ancient DNA from NE Europe previous to the CW culture, so it's hard to say how they were before CW.
- We don't know how the CHG got to Europe (we don't really know well how they got to the steppe either). And the only samples we have are from much before these events. They might not be pure Kotias by that time.
- Archaeology. People who know the details better agree that it's far from clear such mass migration.

So a simplistic equation based on some samples from the steppe and some samples from Germany, and saying 50% population replacement from the steppe is just... simplistic. We have to have many more data points to know what happened and how.

We can't ignore the DNA we have. But we can't ignore that we lack it from many relevant places either. And we can't ignore other data, or to pretend that we know what we don't (like uniparental markers, which are still of unknown origin).

All these questions (or most) will be answered next year. So let's stay open minded and keep a balanced debate. None of us *know* the answers (if someone thinks he does, he's just fooling himself).

Davidski said...

Did you forget that there's R1a in Khvalynsk, Potapovka, Poltavka and Srubnaya? These are all steppe cultures, with Khvalynsk older than Corded Ware and Poltavka as old as Corded Ware.

Semantic arguments like "there's no Z282" on the steppe are a waste of time considering all the data.

The Corded Ware population is indeed very closely related to Khvalynsk and Yamnaya all the way from the Samara and Kalmykia. There's no way around this. It's not some sort of coincidence.

Alberto said...

The steppe samples without European admixture (using Anatolian farmer as a European marker in this case) only have very old and therefor irrelevant subclades of R1a. Coincidence or not, it doesn't tell us much.

The steppe people with modern subclades of R1a have Anatolian Farmer admixture, probably meaning they they moved *to* the steppe from the west (or from Anatolia itself? We don't really know). And they are Z93, which never went to Europe from the steppe.

The R1b in the steppe without CHG admixture is also very old and irrelevant. The ones with CHG admixture are not even ancestral to European R1b: Yamnaya R1b does not even *exist* in Europe, and European R1b does not *exist* in the steppe.

So where are the mass migrations of these steppe cultures to Europe?

So no, I'm not forgetting anything. The data is what it is: no more, no less. It leaves many questions and possibilities. And the trend so far is that the more samples we get, the more sense things make. So I think this trend will continue.

Davidski said...

The steppe samples without European admixture (using Anatolian farmer as a European marker in this case) only have very old and therefor irrelevant subclades of R1a. Coincidence or not, it doesn't tell us much.

At least you're making some progress now. Yes, there's R1a on the steppe.

But I told you already that it's not a coincidence. Corded Ware are essentially a steppe population living in Central Europe.

This isn't just my own conclusion. It's the conclusion of two teams of scientists who put out two major papers on the topic.

Or from Anatolia itself? We don't really know.

Yep, no R1a on the steppe, but plenty in Neolithic Anatolia. Oh yeah, and also in Bronze Age Armenia.

Haha. Holy shit. Maybe cut down on the booze this Christmas.

Balaji said...

Merry Christmas to all!

Very interesting comments from Alberto, FrankN, Davidski, Kurti and others. I tend to agree with Alberto that extent of replacement in Europe and the influence of the Steppe there has been overestimated. No doubt the CW people are related to the Steppe. But the steppe influence was soon diluted. Moreover while the Yamnaya had access to metal tools through trade, they had not mastered metallurgy. The CW people were using stone tools.

I agree with Davidski that people from the Steppe did have considerable influence in the Pamirs. The evidence that he presented is persuasive.

D(Kalash,Tajik_Shugnan)(Kotias,Loschbour)=0.0149 z=3.874

I am glad to see Davidski writing that the influence drops “to nothing deeper in India”. Where I disagree with Davidski is in his interpretation of the following plot.

https://drive.google.com/file/d/0B9o3EYTdM8lQS2FoNnl1RkNtYTg/view?pref=2&pli=1

This plot does show a relationship between Kalash and Sintashta but not because, as Davidski believes, that Kalash has ancestry from Sintashta. Instead I believe that Sintashta has ancestry from South Asia. I agree with Davidski that even Kalash has too much ASI to have been ancestral to Sintashta. But this can be explained by the fact that much of the ASI in what is now Pakistan got there only within the last few hundred years. Both ANI and ASI were present in the Indian Subcontinent since the Pleistocene. ANI was in the western part of the Subcontinent and ASI in the eastern part and were kept apart by the Thar Desert and by dense forests until a few thousand years ago when agriculturalists from the Harappan civilization started moving east. Most of the ASI in Pakistan got there from India after the coming of the Islamic slave trade that took millions of non-Muslims from the east to the west.

We can all look forward to having these issues resolved in the new year.

Davidski said...

Balaji,

Sintashta is obviously a mixture of EHG, CHG and Anatolian Neolithic, and can also be described as Yamnaya/EEF.

So how can it be from South Asia?

Nirjhar007 said...

Balaji,
It will be solved in 2016. The way the things are going, we can truly expect, that we will get the full picture.

Kurti said...

I see someone here insisting on one theory while quite some other posters here try to convince him the opposite, well that is actually good staying to his own opinion. But the other bigger problem I see, is him ignoring virtually ~70% of all the arguments brought up by the other posters and holding on a single straw.

Not a single comment wa


Dave, Renrew is not of my interest because I am not on the Anatolian early farmer hypothesis so why are you mentioning him.

The Indo_Iranian core (or what we call East Iranic branch and Cimmerians) might be from the Steppes, but denying the Kura-Araxes, Hittite or Mitani connection, the fact that the whole West Iranic branch is genetically too distinct to be directly from Sintashta and going by acient records and legends came either from Kura-Araxes or Yaz culture.

Also did you completely ignore Horse traces in Kura-Araxes by 3300 BC, predating Sintashta by 1200 years.

You ignored that the earliest Kurgans are found in the Transcaucasus (Leyla Tepe, Maikop, Kura-Araxes).

And well if you insist on the wive theory so be it. But let me tell you something. the MOTHERtongue is actually teached by the mother. I have witnessed at first hand how in mixed marriages the children were speaking the tongue of the mother, rather than that of the father whom they saw only half the time. I encoutnered a Turkish dude with his obviously East European looking wive and two children.

And I was quite stunned to see that these children were speaking fluently Russian with their mother but could not exchange a single word in Turkish with their father who tried hard.

I have been witness of these kind of things too often. We are quite underestimating the influence the mothers have to the children they are keeping an eye on almost 24/7.

Not to get me wrong I am still absolutely against the CHG wive theory because it doesn't make sense in my eyes and with the archeological findings and cultural impact from South into Yamna. Just trying to make clear that believing all the CHG influx came via wives, doesn't bring you closer to an "EHG origin of PIE".

Happy holidays to the Christians posters here.

Unknown said...

You guys that think they have South Asian ancestry must be high. CHG clearly goes down, along with EHG. It's Anatolian AND WHG that increase. It is admixture from Central Europe. Not South Asia. How the hell do you think they look like Modern Europeans? Common sense here.

Unknown said...

Sintashta is about identical to Bell Beakers. Are you going to say Bell Beaker is South Asian. There's no damn Asian admixture, AT ALL.

Iranocentrist said...

What if its the CHG admixture hiding their emmediate west asian or Central asian origins?

Davidski said...

Some of you guys are so desperate now you actually sound insane.

FrankN said...

Since it seems to be at the heart of this blog, let me add my two cents on the PIE homeland here. Hopefully, it cools down discussion…

1. PIE never existed. It is a theoretical construct by linguists. What existed was a prehistoric dialect continuum, or a cluster of closely related languages, the unifying element of which has been abstracted into PIE.

2. PIE formed as a Creole. To linguists, a Creole is a specific kind of hybrid language, where Language A contributes most of the vocabulary, while grammar/morphology and a substantial part of phonology stem from Language B. It develops in a stratified society where status is related to ethno-linguistic background.
Speakers of language B, usually predominantly male, constitute the socio-political elite, while language A, usually prevailing among women, is in use in the families. Kurti has illustrated how the mother tongue A comes to dominate the general lexicon. Even if Language B is used for some time in everyday's life, it will quickly disappear with the 2nd, 3rd, 4th etc. generation of A-speaking women transferring their mother tongue to the children.
However, mastering Language B, the "elite language", is indicator of social status, proof of descent from the "founding" (or conquering) fathers. Thus, language B's grammar and phonology (mostly) prevails. In fact, the more complicated Language B is in this respect, and the more it differs from Language A, the better, because it makes gauging social status easier. This implies that mothers, while transferring Language A vocabulary, will pay attention to their sons using Language B's "correct" grammar and accent.
An illustration is Romance languages, which have essentially preserved (Celto-)Roman vocabulary (A), but whose grammar has been thoroughly Germanized (B): Use of articles, personal pronouns, tense-building with auxiliary verbs, loss of cases/ case markers, etc. Lat. vidi "I have seen” - “J’ai vue".
[The “un-IEness” of Germanic, btw. demonstrates it as latecomer to the IE family, possibly under Unetice influence during establishment of the Nordic Bronze Age.]

3. PIE is partly Caucasian: Dave some time ago linked to a Bouckaert paper that claimed relation between PIE and Proto North Caucasian (PNC). To me,, the phonological and morphological arguments seemed convincing (but I am not a trained linguist, this isn't expert judgment). Then came the wordlists, and I thought: Crap! Pastoralist terminology was fine, also kinship terms, but that was about it - not a single convincing PNC – PIE relation for body parts, verbs of motion, the physical world etc.
But the longer I thought about it, the more I realized that this is exactly what should be expected from a PNC-based Creole. Kinship terms are obviously superstrate, as is pastoralist terminology (if not, it is at least related to a typical male domain). Plus the grammar, plus significant influence on phonology. Everything else, i.e. most of the lexicon, found in a PNC Creole should be from Language A, the mother tongue.

4. PIE is more than Caucasian : Caucasian men taking Steppe woman isn't exactly what has been put forward here as scenario. And it doesn't need to be. Because in all likelihood the Proto-Caucasian spoken 5-7 millennia ago itself already was a Creole or some other kind of hybrid, evolved in millennia of CHG/EHG (EEF?) interaction after the end of the ice age. And whatever was the Language A it hybridicised with - that language should also not have been "pure" Cantabrian or Altaic LGM refugia speech, but equally the result of CHG-EEF-EHG interaction.
My guess for Language A goes towards Pre-Proto-Germanic (more precisely the pre-IE substrate conserved in Germanic), but we might also have some Pre-Proto-Uralic here (or actually one day have linguists demonstrating that both were closely related). Anyway, sorting that out is something for linguists, and seems to be a pretty challenging task.

FrankN said...

5. PIE is in the admix : My view is that PIE has developed in a process of multiple hybridisations between languages spoken by WHG, EEF, CHG and EHG. Most likely not once, but several times, resulting in a cluster of languages that, for shared roots, were close enough to each other to provide linguists with the material they have abstracted PIE from. Interaction between the speakers of such early IE languages should have provided for further homogenization.

6. The PIE homeland is the admixture zone: The Steppe? Possibly (though Yamnaya lacks the WHG portion that is a constituting element of the PIE admix). The South Baltic shores? Looks like, if we accept CW people as speaking a PIE-related language. Romania /Moldova/ Carpathians? May well be – let’s wait for aDNA from there. In any case, if they spoke some early IE, it should have differed from the CW version in incorporating more EEF and less EHG vocabulary. Transcaucasia? Why not, if WHG/EHG made it there early.
My main point here is that it isn’t about either/or. The world’s largest language family must have originated in a wide area, and among a sizeable population to rise to the relevance It has today. We are starting to understand the underlying mechanism, namely several populations that during the ice age had been isolated from each other and scattered across Western Eurasia from India to Europe gradually merging with each other (sometimes peacefully, sometimes not so) and creating a (North-)Indo-Caucaso-European genetic mix that also homogenized their languages. When another place shows up demonstrating such admix, let’s add it to the homeland.

7. There is no Pre-PIE. I am curious to learn where R1a/b found refuge during the Ice Age, and whether it was the same or two separate places. I am equally curious to learn more about the paths they took to show up where we find them today. The same applies to J2, and other yDNA (and, of course also mtDNA) haplogroups. But if R1a is genetically demonstrated to be EHG, which seems to be the case, this is hardly relevant to the PIE question, because PIE wasn’t a EHG language, but developed as a CHG-EHG-WHG-EEF Creole.
O.k., the R1a / EHG question isn’t completely irrelevant. It holds one quarter of the answer on how the admix evolved. And that quarter is far less known than the EEF story, and also somewhat more obscure than the WHG and CHG parts. So, worthwhile of digging deeper. But a quarter of an answer shouldn’t be taken for the full answer.

Finally, a delayed whish for Christmas, or a proposal for a New Year’s resolution: Could we please withhold inflationary use of the word “kurgan” as long as the Mississippi Mound Builders haven’t been demonstrated to have spoken IE? I mean “crouched burial, eastward oriented, ochre-painted, in a wooden cist covered by an earth tumulus” may look a bit clumsy as formulation, but would help me to tell if the apple is an apple or an orange.

Davidski said...

First of all, PIE was not a creole.

Secondly, the Kurgan people of the Middle Bronze Age did speak Indo-European, because they are the only plausible link between Eastern Europe and South Asia within a reasonable timeframe.

Vedic rituals described in the Rig Veda can be seen in the Kurgan burials on the Trans-Urals steppe. And the people buried there belonged to R1a-Z93, which today peaks in frequency in South Central Asia and is the sister clade of the Balto-Slavic-Norse R1a-Z282.

It's very difficult to wriggle out of this considering the archaeological, genetic and linguistic data.

Kurti said...

@FrankN said

"1. PIE never existed. It is a theoretical construct by linguists. What existed was a prehistoric dialect continuum, or a cluster of closely related languages, the unifying element of which has been abstracted into PIE."

Thank you, see my post above. PIE is just a construct by linguists to explain linguistic similarities which can only be explained by some close connection and some shared ancestry. IMO Indo Europeans can not be modeled by just one culture or "Proto people". But rather we are dealing with an network of "Proto" Indo European culturs.

Rob said...

@ Kurti

In my view to some of yur comments:

"In my own theory, Maykop, Kura-Araxes, cucuteni-tripolye, Sintashta CW and probably also Leyla Tepe are as much "PIE" cultures as Yamna."

What we're seeing is a elements of a shared *ideological* language amongst Copper Age elites; construction of mounds, burials with prestige copper objects, horses, wagons, etc over an expansive region. Yes, IE was probably one language amongst these areas, but Im sure there were others. That is, this was not a monolithic lingual block

"Many scientists are actually supporting a Maykop/Kura-Araxes origin of Hittites/Anatolian Indo Europeans."

This is a large gap: 3800 BC (start of Majkop) and 1800 BC (earliest attestation of Hittites). Simply, there are too many tenuous assumptions to make this hold true, and we know copper age and early BA polities & ethnies were too labile to maintain themselves for 2000 years. Nor can we trust later myths which are simple archaicizing topoi used by ancient authors everywhere (ie linking contemporary people with bygone peoples).

The Majkop origins of Anatolian IE was entertained in the 80s, when the age of the Majkop kurgans had not been confirmed to be so old. It was believed many metal finds from Anatolia are similar to Majkop. But again, the problem is the chronological discrepancy, as the Anatolian objects date from the late 3rd Millenium, and are probably just chance resemblances

Of course, this is not to say that Majkop did not have a formative role in IE languages, I just wouldn't (yet) call it proto-Anatolian or proto-IA.

Kurti said...

Hi Rob

"This is a large gap: 3800 BC (start of Majkop) and 1800 BC (earliest attestation of Hittites). "


Well the timeframe fits actually perfectly with the collapse of the Kura-Araxes culture.

Just ~2000 BC when Kura-Araxes culture collapse we have Indo European cultures such as the Hittites and Mitanni in Anatolia anbd Mesopotamia.

And once again since the Rig Veda stuff and Sintashta was brought up.

I didn't said a single time that Sintashta is not the ultimate source of Indo_Aryans. In fact I said BMAC is the source of Indo_Aryans which is supported by archeologists. And BMAC is basically in the sphere of Sintashta influence.

Sintashta => BMAC => Indo_Aryans.

What we are discussing here is the origin of the Indo European language family as a whole and imo, we are dealing with a network of Indo European cultures rather than one single Indo European culture.

Mitanni for example is too early to have reached the region via Sintashta. And Mitanni seems rather to be an archaic form of Indo_Iranian than really Indo_Aryans. It seems to be a Cimmerian like case. Cimmerians as most likely an early outgroup of Indo_Iranians originated from the Srubna culture and not Sintashta. So if we have proofs of Iranian tribes outside the Sintashta frame, why shouldn't there be other Indo_Iranian tribes with the same case? For example the Mitanni(who most likely are related to the Medes) or the whole West Iranic branch?

In ancient Achamaenid texts only the former Kura-Araxes territory and Yaz culture territory are given any significance in the question of their origin.

Rob said...

Not my area of expertise, but i thought that the gods names and couloirs in the Mitani texts are actually archaic I-A, not proto-Iranian ?

Kurti said...

Other interesting notes which seem to be forgotten always.

We still don't know what kind of language Hurrians really spoke. I assume them to have spoken a language that has been a important factor in the whole Indo European language family.

Many things about the Hurrian culture scream Indo European without any sign of being later introduction. For example the Hurrian Sky God Teshub is basically the proto type of Zeus, Hattian Taru, Hittite Tarhun, Celtic Taranis and Vking Thor.

How can such a connection exist if not for shared ancient etymology?

I don't know of any contact betwen Celts, Vikings and Hurrians and Hattians.

These kind of things make me believe that we are dealing with a network of Indo European and closely related (non Indo European) cultures and with FrankNs words a "Creole" of language family. This would explain why we have early Indo European languages which have more vocabulary with farming (Tocharian) while others seem to completely lack it.

Kurti said...

@Rob
Not my area of expertise, but i thought that the gods names and couloirs in the Mitani texts are actually archaic I-A, not proto-Iranian ?

All the names in the Mitani texts are basically Indo_Iranian gods that were more relevant in Indo_Aryan than Iranian, what cause some experts to think that we are dealing here with archaic Indo_Aryan but the matter of fact is actually the more archaic the Indo_Iranian language is, the more it would appear closer to Indo_Aryan simply out of the fact that Indo_Aryan itself stayed significantly more archaic in it's whole form. It is significantly closer to Proto Indo_Iranian than the Iranic branch. Which is also the reason why other Indo European languages such as Lithuanian show closer relationship to Indo_Aryan than Iranic, despite geographically the opposite should be the case.


However even here many people make the mistake to believe that the "Indo_Aryan" origin of Mitanni is widely accepted when in fact as I mentioned above it is so archaic that there is still debate about wether it is some sort of archaic undivided Indo_Iranian or archaic Indo_Aryan. Which tells much.

And as I said naturally the more archaic the more it will appear closer to Indo_Aryan.

I quote from Wikipedia.

"A treatise on the training of chariot horses by Kikkuli contains a number of Indo-Aryan glosses.[9] Kammenhuber (1968) suggested that this vocabulary was derived from the still undivided Indo-Iranian language,[9][10] but Mayrhofer (1974) has shown that specifically Indo-Aryan features are present.[11]"


So the debate is still there. what can only mean that it is very archaic.

Rob said...

That is what I thought , and I think we are in general agreement in chronology and basic geography

Davidski said...

We still don't know what kind of language Hurrians really spoke. I assume them to have spoken a language that has been a important factor in the whole Indo European language family.

Get off the medication ASAP.

https://en.wikipedia.org/wiki/Hurrian_language

Btw, Mittani were Hurrians with an Indo-Aryan ruling class.

Kurti said...

Dave contact mewhen you actually have understand the difference between a theory and a solid proof.

If you actually took the time to read a single sentence written by any of the posters here your comment would have been obsolet, cause no one said Hurrian is Indo European.

My comment still remains , we still don't know to which language family Hurrian belonged. There are merely speculations.

Davidski said...

Hurrian texts are solid proof that their Indo-European influences came from their Indo-European ruling class. If that doesn't meet your burden of proof, then nothing will, because time travel is impossible. But no one cares.

Rob said...

Dave

Maybe I'm mistaken but I think Kurti accepts that IEs were the ruling elite, he's rather pointing out that the classification of Hurrian base language itself is still a matter of debate, incl it's supposed Caucasian and Urartian affinities

Davidski said...

I don't know what the hell he's rambling on about. Can you make out what this means?

We still don't know what kind of language Hurrians really spoke. I assume them to have spoken a language that has been a important factor in the whole Indo European language family.

WTF?

Balaji said...

Davidski, Chad Rohlfsen,

I agree with you that Sintashta can be modeled as Yamnaya and EEF. That is why as Alberto pointed out, it surprising to see the high affinity of Sintashta to Kalash since Kalash has very little or no WHG or EEF, The Smarter Bear plot shows how little WHG Kalash has.

https://drive.google.com/file/d/0B9o3EYTdM8lQb1R2MDJmS2h0Nk0/view?pli=1

As regards EEF, see the following from Chad Rohlfsen

result: Anatolia_Neolithic1 Kotias Kalash Primate_Gorilla -0.0132 -3.571 22044 22632 454082
result: Anatolia_Neolithic1 Kotias Georgian Primate_Gorilla -0.0011 -0.312 22532 22581 454082

Kalash cannot have much ancestry from Sintashta. Therefore the connection must be due to the ancestors of Kalash contributing ancestry to Sintastha and not the other way around.

Davidski, can you place Anatolia_Neolithic1 on your Smarter Bear plot? It should be below and to the left of Starcevo_EN.

Davidski said...

Balaji,

Here you go. I hope this helps.

https://drive.google.com/file/d/0B9o3EYTdM8lQWmhKWk14bFhkcjg/view?usp=sharing

Nirjhar007 said...

Vedic rituals described in the Rig Veda can be seen in the Kurgan burials on the Trans-Urals steppe.
Yes , i have read plenty of times while studying Rigveda, the most popular one was to cut a dudes head and replacing it by a horse.
Kurti,
Mitanni seems rather to be an archaic form of Indo_Iranian than really Indo_Aryans.
Actually if we notice the nature of the words they are of typical Avestan or Iranian type, imo the Mitanni dudes were linguistically more closer to Iranians but in case of religion were to Indians, i suspect its because they were unaffected of the reformation of Zoroaster, which probably happened around the same period but happened around Bactria..
Indo-Iranians have their roots in SC Asia, the Steppe lands were the play grounds of the Pre-Scythian (Andronovan ) folks, they were a part of the IIr family.

Nirjhar007 said...

Kurti,
My comment still remains , we still don't know to which language family Hurrian belonged. There are merely speculations.
Hurrian do show sharing of common roots with IE, i'm at the moment busy on elaborating the IE-Sumerian but will focus next on Hurro-Urartian, Caucasian etc.

Grey said...

FrankN

1) "By 3000 BC, however, there are strong signs of reduced settlement intensity across Southern Scandinavia and parts of Central Europe. Reasons are not fully understood yet."

"He told me that he didn't need aDNA to figure out that there was a major population turnover around 3000 BC - a look at the skulls in his collection would do."

2) "However - to make it equally clear: There is nothing in the German archeological record suggesting anything like chariot-driving cattle herders conquering locals to bring about that change."

Signs of reduced settlement intensity would be consistent with displacement by a non-settled mobile pastoral population with their horses and cattle.

http://grthm.natt.org.uk/images/02.jpg

postneo said...

@david.

you were right I did not read the title properly, "three distinct genetic...". But the agreement ends there. if more genetically distinct isolatable poles/components are identified we could easily say 4 or 5. Plus we don't know any of the languages spoken. You are trying to model an assumption based on ancient dna on one side and only modern dna on the other.

"Vedic rituals described in the Rig Veda can be seen in the Kurgan burials on the Trans-Urals steppe"

Thats wrong and second hand knowledge. Please read about these rituals or observe first hand to evaluate. You will find that the resemblances are stray and non mainstream. Out of a ton of extremely elaborate ritualistic descriptions a few have indirect resemblance. This is like saying that you have deciphered a script/language because two characters match that of a known script.

First of all vedic ritual does not involve burials or horses chariots or anything. Perhaps you can say there is some indirect resemblance because both refer to horses. You have to show these resemblances are beyond coincidence and they are actually the same people.

Also the freakish/anomalous preservation of vedic skews the perception of IE as being horse centric. Its not necessarily true all IE speakers were horse centric. A shared indo european trait is one of indra(the pre-eminent vedic god) / apollo fighting a snake with a weapon made of sinew/bone/foam ... something whitish. This is non vedic and from the puranas yet resembles greek myth. When evaluating this kind of kinship, the first order of business is to make sure whether this is even an exclusively indo-european trait before jumping to a specific vedic affiliation.

David Anthony's articles are well written but biased. His wheel line is a good hypotheses, but he is guilty of overly humanizing his favorite Steppe subjects .. they were just like us ..they threw parties ...youv'e got to fund this....etc..

Chad said...

Balaji,

You're not understanding the stats and looking at it backwards.

That stat just means that the Kalash have much more CHG than Anatolian. It doesn't mean that they have little to no Anatolian. Some of that affinity is going to be due to the slight shift towards EHG from Kotias, which the Kalash seem to have a good amount of. So, don't put the cart before the horse too much. There is no South Asian ancestry in Yamnaya, Sintashta, or Europeans. Moving on...

FrankN said...

@Grey: "Signs of reduced settlement intensity would be consistent with displacement by a non-settled mobile pastoral population."
Yes, they would. But that is not how things worked. Dolmen (partly built over long barrows that were built over shell middens), erected during the 3600-3300 population boom, continued to be used into the BA Tumulus Culture, sometimes into La Tene. (Semi-)continuous use of the same burial site over millennia isn't really speaking in favour of population displacement.

The common scenario, documented for the LBK-Rƶssen-Michelsberg transitions, was decrease of the "old" population, which subsequently concentrated in the most suitable sites for their economic model (i.e loess soils in case of LBK, wetlands with more pastoralist Rƶssen/ Michelsberg), while "newcomers" gradually occupied the emptied areas.

Dolmen "recycling" is a main part of the insecurity about population numbers. There was an estimated 150,000 tumuli ("HĆ¼nengrƤber") in Northern Germany, Denmark and Skane (Polish archeologists are still counting). While there is construction gap after 3300 BC (Germany)/ 3100 BC (Denmark/ Skane), existing dolmen continued to be used, regularly cleared after a few generations. Later cultures eventually built additional ones if the existing weren't sufficient anymore. This means that each of the tumuli that still exist would need to be opened in order to check to which building phase it belongs, and whether there is indication for continuous use, or some temporary gap. Fig. 1 in the link gives an idea about the monuments that so far have been safely assigned to TRB.

http://www.academia.edu/1346797/Megaliths_and_Identities._The_earliest_monuments_in_Europe_-_architecture_and_social_structures_5000-3000_cal_BC_

Depending on the estimated household size (6 seems to be a fair guess) and the number of tumuli already dating back to TRB (50-150 thsd.), you get population estimates anywhere between 300 and 900 thsd. Quite a margin, but in any case to numerous to be easily displaced.

Pollen diagrams from Holstein show a strong decrease in cereal-related indicators by 3000 BC (Fig. 6). But the landscape stays open, and shrub pollen (->hazelnut!) increase markedly. Apparently a shift back from Neolithic farming to Ertebolle-type aquatic foraging (climate, soil exhaustion); additional entrance by dairy pastoralists looks likely.

The Berlin periphery has been thoroughly studied over the last decades due to lots of post-unification transport and estate development. There, strong settlement increase, with cultural linkage to TRB_N, took place around 3500 BC. A few centuries later follows further intensification (overforming, not replacement) by GAC, leading to a settlement structure of 1 house every 400m along the watercourses and lake shores. No population estimate is available, but there are a lot of watercourses and lakes around Berlin. Soils are sandy and poor, hardly suited for farming, so settlers seem to have used a mix of aquatic foraging (isotopic analysis shows high fish content in the food) and dairying. That dense settlement structure remained essentially unchanged for the next 1400 years, i.e. into the Bronze Age (link, p.12, in German).

http://www.dirk-schimmelpfennig.de/ag_neo/pdf/AG-Neolithikum_2013_Programm_Luebeck.pdf

Another abstract (p. 4) presents recent discovery of a late 3rd mill. pile-dwelling site in Veska (Vologda, RUS, Combed Ware), together with possible fish-trap remains. I haven't found any detailed description yet (maybe in Russian?). The following publication helps to at least put it into context, and towards the end provides a good overview on the cultural sequence and pottery flows in Mesolithic NE Europe.
http://arheologija.ff.uni-lj.si/documenta/pdf39/39_2.pdf

Iranocentrist said...

Can anyone find any info on the physical appearance of the Mitani elite?

Grey said...

FrankN

"The common scenario, documented for the LBK-Rƶssen-Michelsberg transitions, was decrease of the "old" population, which subsequently concentrated in the most suitable sites for their economic model (i.e loess soils in case of LBK, wetlands with more pastoralist Rƶssen/ Michelsberg), while "newcomers" gradually occupied the emptied areas."

That works.

FrankN said...

Grey:

Here is an older analysis (2003) of the EN-MN transitions in the Elbe-Saale region. Its in German, but with English language abstract and map/diagram legends. A look at the sequence of maps is making the general pattern clear.
Don't ask me about the intermitting cultures of Pass Linear Pottery and Gatersleben - I haven't yet understood their specifics (I am not sure anyone has). The map on the Pass Linear Pottery suggests a somewhat "southern" origin (Bavaria/ Austria?), they don't appear to have spread via Bohemia. Rƶssen originated west of the Rhine and is also poorly researched. I found a mid-1980s presentation contemplating lack of Rƶssen research since the 1930s, plus a few hints in conference abstracts on "Rƶssen spatial patterns, probably also economic strategies, differing from LBK." But fortunately we have the mtDNA from Brotherton/Haak 2013 showing them full of "Basque" subgroups of mtDNA H. A tentative link is made to the La Hoguette hunter-pastoralists from the Western Alps/ Middle Rhone.

http://www.jungsteinsite.uni-kiel.de/pdf/2003_ostritz.pdf

MĆ¼ller's team in Kiel has made quite some attempts to sort out the MN, leading into what he has called the "Central German Neolithic Multicultural Approach". The map in Fig. 11 of the following English language paper makes you understand how that concept emerged. MĆ¼ller's chronology (Fig. 10) is now generally accepted. If you take a look at Fig. 1, you will grab the overall high settlement density in the area. The NE corner is the Berlin area I wrote about in my previous comment. Where settlement appears to fade out towards the map corner lies Berlin proper - most likely no less densely settled during MN/LN, just far less excavated.

https://journals.uair.arizona.edu/index.php/radiocarbon/article/viewFile/3529/3044

These mixed, multicultural patterns, which don't show up on any of the WP maps usually consulted, need to be kept in mind when interpreting data. Haak started his career in the Elbe-Saale region and pays a lot of diligence to assigning each DNA sample the proper cultural affiliation. The fact that there is a clear genetic change from Baalberge to CW thus means a wave of new immigration. But it doesn't mean the Baalberge, Bernburg, GAC people dying out or being displaced, nor does it mean CW DNA being representative of the total population in the Middle Elbe region after 2800 BC. More MN-like aDNA reappearing with Bell Beakers demonstrates CW only being part of the region's demography - though a quite significant part, as its admix left strong traces in today's population.

Let me finally say that while decline-concentration-new settlement of emptied areas is the prevailing pattern, there is indication of some "non-consensual" expansion and displacement. These are the cases I have come across so far:
1.) Violence among Michelsberg culture enclosures/ fortresses: Destruction of the Talheim site, substantial burning of the heavily fortified eponymous Michaelsberg site, both near the Upper Rhine.
2. Polish GAC cutting off the Baden settlement zone around Cracow from their traditional flint sources in NW Ucraine.
3. The Bernburg culture (Michelsberg-Nordic FB-GAC) by 3100 BC destroying SalzmĆ¼nde proper (Baden-influenced) including their graves, and subsequently completely replacing SalzmĆ¼nde A/B layers with own material in Thuringia and the southern part of the Middle Elbe region.
4. The CW population in Eulau being massacred, most likely by the Schƶnfeld culture (Harz to Berlin area, FB with GAC overforming).

To me, this rather points to the "bad guys" coming from the (North-)West and taking land in proto-Celtic-Viking manner, than to a steppe invasion. Should additional evidence show up, I am prepared to review this assessment which is admittedly nothing more than extrapolation of scarce and sketchy evidence.

Rob said...

Frank

Interesting insights to the German Neolithic, thanks as it is sometimes hard to come across anglophone literature on this

Is your theory that PiE and R1b came from north or western Germany ?

FrankN said...

@Rob: "Is your theory that PiE and R1b came from north or western Germany ?"
That would be clearly oversimplifying.
On R1b I am pretty agnostic and waiting for further aDNA to come. I am anyway getting lost among all the subgroups. A rather unsubstantiated theory, more a guess than anything else, would be some small admixture carried along by EEF to both Germany and Iberia, and by some chance becoming dominant (IIRC there has been some R-like stuff that couldn't be analysed further downwards showing up in German EN).
My pet theory is that R had developed some kind of early immunity against the Plague, posssibly already during the Ice Age, which helped it becoming dominant in Europe. Mortality patterns during the Black Death are pretty strange, with certain regions being devastated (Tuscany, South England, NW Germany, Norway, Iceland), while nearby regions were hardly hit at all (e.g. Lombardy, Basque country, Bohemia, Silesia). Wales apparently also did reasonably well in comparison to Sussex and East Anglia; Medieval Swedish graveyards show yDNA I1 much more dominant than it is today.

On PIE I have made my point - it developed as CHG-EHG-WHG-EEF linguistic hybrid, more than once and in more than one variant. The WHG-heavy portion (with EEF and possibly EHG admix) I could well imagine to have originated with Ertebolle & related Baltic cultures. There is a clear and strong south-eastward expansion out of the post-Ertebolle Nordic Funnelbeaker zone. In NE Germany, it fused with Michelsberg into Bernburg and then swept over Baden-derived SalzmĆ¼nde. GAC seems to have originated from a similar fusion, here FB with Epi-Lengyel and possibly some Narwa Culture elements. Both may in fact have included quite some CHG, if I interpret the spread of millet farming and certain pottery traditions & recipes correctly - let's wait for respective aDNA.
The GAC expansion towards the Black Sea, and into Yamna territory rather than the other way around, is archeologically well documented. If they were on the move already, they may also have continued further towards Transcaucasia - that would be a plausible explanation for SE Carpathian, Armenian and Central Anatolian domestic pig DNA being completely replaced by such of Central European origin.

So - I don't see PIE coming from NW Germany. But I see an expansionist process starting there (actually already in the Paris Basin, possibly even with La Hoguette on the middle Rhone), and gaining further momentum in Southern Scandinavia and Poland, which substantially contributed to linguistic hybridisation north of the Pontic and Caspian Seas, and had sufficient impetus to reach into Anatolia, towards the Altai and beyond (obviously in both cases picking up more CHG/EHG loaded DNA on the way).

What I am still wondering about is whether we had a kind of 4-way lane Bug-Dniepr highway (post-FB/GAC moving SE, and CHG simultaneously moving NW), or, as the spread of pottery seems to suggest, CHG already earlier had taken the easterly Volga route, merged with EHG in the NW Russian Forest cultures, to enter Central Europe from the East Baltic when half of FB/GAC had just left for the Pontic/ Caspian shores.

Grey said...

FrankN

"The fact that there is a clear genetic change from Baalberge to CW thus means a wave of new immigration. But it doesn't mean the Baalberge, Bernburg, GAC people dying out or being displaced"

Fair enough. The total farmer range shrinking due to changed conditions and pastoralists moving onto the abandoned part of the land is clearly plausible also.

When a lack of settlement evidence for pastoralists is mentioned i like to throw in a what if. What if they lived in their wagons?

FrankN said...

"What if they lived in their wagons?"
Might have worked in some areas. In the Harz mountains probably less so. Central Lithuania is quite forested, from what I remember. Forest grazing provides some space, but not neccesarily the headroom a wagon requires. Coastal and Upper Elbe marshes, flooded from time to time, aren't prime wagonracing terrain either.

http://www.neuepresse.de/var/storage/images/np/nachrichten/kultur/uebersicht/heavy-metal-in-wacken-laut-friedlich-und-schlammig/23676074-7-ger-DE/Heavy-Metal-in-Wacken-Laut-friedlich-und-schlammig_ArtikelQuer.jpg

So - no, I don't think so.

Most likely, unlike the LBK people, they settled where no motorway has recently been constructed*, and/or on terrain prone to erosion or flooding. Lighter, temporary hut constructions are anyway difficult to recover (Until a few years ago, it was believed Ertebolle people had no permanent settlements. In the meantime, dozens of them have been excavated some 5-7m below the current Baltic Sea level).

*) Note that most of the Elbe-Saale aDNA we have is from just a handful of highway and railway construction projects over an area measuring roughly 60x60 km. Prime transport corridors today, as in the Neolithic, and well within the typical range for exogenous marriage. If somebody has the time - it might be interesting to check in how many of the EN-MN-LN comparisons we are actually dealing with direct relatives..

Rob said...

Krefter , Mark

I think Mark is suggesting its still theoretically possible that R1b is somehow a MN west European marker, and the 'steppic' ancestry arrived via some admixture with CWC and female exogamy. I guess this is not parasiminous but can only be disproven by finding L51 in east-central Europe in 2016.

Rob said...

@ Franky

Hi Franks, thanks for your reply. I was being slightly cheeky, for i knew you werent really insinuating that IE came from NW Germany. I;ll reply to a few of your points:

* 'On R1b I am pretty agnostic and waiting for further aDNA to come. I am anyway getting lost among all the subgroups. A rather unsubstantiated theory, more a guess than anything else, would be some small admixture carried along by EEF to both Germany and Iberia, and by some chance becoming dominant (IIRC there has been some R-like stuff that couldn't be analysed further downwards showing up in German EN). "

- yes, who knows ? Oviusly the dominant theory places it on western steppe, but Im leaning toward the east Carpathian region/ northern Balkans ?

* 'My pet theory is that R had developed some kind of early immunity against the Plague, posssibly already during the Ice Age, which helped it becoming dominant in Europe. Mortality patterns during the Black Death are pretty strange, with certain regions being devastated (Tuscany, South England, NW Germany, Norway, Iceland), while nearby regions were hardly hit at all (e.g. Lombardy, Basque country, Bohemia, Silesia). Wales apparently also did reasonably well in comparison to Sussex and East Anglia; Medieval Swedish graveyards show yDNA I1 much more dominant than it is today."

- Plague resistence developed in LGM ? Hhmm

but I agree that studies from medieval populations show that they are more diverse, so the plague must have had some effect even if Europe already numbered in the millions at that stage


* "On PIE I have made my point - it developed as CHG-EHG-WHG-EEF linguistic hybrid, more than once and in more than one variant. The WHG-heavy portion (with EEF and possibly EHG admix) I could well imagine to have originated with Ertebolle & related Baltic cultures. There is a clear and strong south-eastward expansion out of the post-Ertebolle Nordic Funnelbeaker zone. In NE Germany, it fused with Michelsberg into Bernburg and then swept over Baden-derived SalzmĆ¼nde. GAC seems to have originated from a similar fusion, here FB with Epi-Lengyel and possibly some Narwa Culture elements. Both may in fact have included quite some CHG, if I interpret the spread of millet farming and certain pottery traditions & recipes correctly - let's wait for respective aDNA. "

I dont always agree with what linguists *think* they conclude, but one thing we have to accept is that PIE isn;t a "hybrid". It's a primary, 'natural' language- but this is not to say that substrate effects and later convergences weren't operant.

* "The GAC expansion towards the Black Sea, and into Yamna territory rather than the other way around, is archeologically well documented. If they were on the move already, they may also have continued further towards Transcaucasia - that would be a plausible explanation for SE Carpathian, Armenian and Central Anatolian domestic pig DNA being completely replaced by such of Central European origin."

Yes there was certainly some GAC movement to SE, but wasn;t this reversed ? Indeed, CWC pottery might have its origins in the east Carpathian / lower Danube region.

* "What I am still wondering about is whether we had a kind of 4-way lane Bug-Dniepr highway (post-FB/GAC moving SE, and CHG simultaneously moving NW), or, as the spread of pottery seems to suggest, CHG already earlier had taken the easterly Volga route, merged with EHG in the NW Russian Forest cultures, to enter Central Europe from the East Baltic when half of FB/GAC had just left for the Pontic/ Caspian shores"

I think you're forgetting about the remaining 75% of Europe - and the main artery of Europe - the Danube. At the moment, it really appears that CHG entered southern Europe, and 50% of that in central Europe, via the Balkans. This will eventually need to be factored into peoples hypotheses. The Bug-Dnieper region is a sideshow to the big picture.

FrankN said...

@Rob:

"PIE isn;t a "hybrid". It's a primary, 'natural' language."
In linguistic terminology, "hybrid" is used for any outcome of language contact.
https://en.wikipedia.org/wiki/Language_contact

There is nothing like a "natural" language outside computer programming. All spoken languages are hybrids and have adopted features from other languages. 99% of the words in the Oxford English Dictionary, e.g., are borrowings. That doesn't make English "unnatural", just demonstrates its intensity of external contact. Question isn't whether there was hybridisation, it is about its nature and intensity.

Many people think of Creoles only as outcomes of contact between Colonial Europeans and "Natives", but there are, e.g., purely African-based Creoles such as Sango. Analysis of Austronesian, which displays a quite homogenised grammar/morphology, but substantially differing vocabularies among the sub-families, is a kind of forerunner in applying the concept of Creoles to historic linguistics.
The link below demonstrates typological features (grammar, morphology) changing at similar rates as lexicons, but unconnected to them. This renders the standard equation "language family = shared lexicon & shared typology" questionable. Note the typological neighborhood network (Fig.1) showing, a.o.,
- French branching from German,
- Finnish (but not Hungarian!) sharing a root with the Western IE cluster,
- a Caucasian-Burushaki-Hindi-Kannada(Drav.)-Turkic cluster, from the root of which also spring Basque (Cauc. side) and Quechua (Turkic side)
- a Japanese-Korean-Burmese branch (with the root shared by KhoeKhoe),
- neighbouring Afro-Asiatic and Eastern Niger-Congo clusters, but Western NC (Yoruba, Sango) located on a distinct branch that is related to a Polynesian-Indonesian- Thai-Vietnamese cluster.
- The Greek-Albanian-Bulgarian-Romanian branch described in the Supp.Mat. is also worth noting.

http://rspb.royalsocietypublishing.org/content/277/1693/2443

Are these clusters "natural"? Clearly, they hold a lot of information on language contact - partly unsurprising and potentially well explainable (e.g. clustering of Eastern NC and Afro-Asiatic should relate to medieval Arab trade with East Africa), in other cases (Yoruba-SEA, Quechua-Turkic-Dravidian) deserving further analysis and follow-up. Their macro-cluster 1 (North Eurasian) seems to be pertinent to the overall genesis of PIE.

As concerns R1b - even more puzzling than the disjunct (paleo-)Eurafrican distribution (Karelia, Iberia, African V88) is the North American occurrence. We have quite an idea by now on the possible scope of contact-induced genetic change. 62.5% R1b in Chippewyan looks far beyond anything expectable from post-columbian contact, which only started in the 19th century. The same applies to the 80% with Ojibwe, for which first European contact is documented for 1640.
An older analysis (2008, link) had identified 19 (out of 63) R1 haplotypes in NE Amerindians that were not shared with Europeans, 5 of them more than 4-8 mutations removed from European variants. In the meantime more European variants should have become known, so the a/m results may be exaggerated or outdated. Unfortunately, I haven't yet seen any follow-up.
http://mbe.oxfordjournals.org/content/23/11/2161.full.pdf+html

There certainly was much more going on across seas than we are yet aware of - a couple of linguistic and genetic signals (e.g. mtDNA X2, dog aDNA, bottlegourd aDNA) point to transatlantic and/or transpolar exchange. To repeat myself: I am curious to see where R1(b) turns up next. And I am pretty certain it will, in addition to Basques and West Africans, include other people not speaking IE.

FrankN said...

@Rob:

"I think you're forgetting about the remaining 75% of Europe - and the main artery of Europe - the Danube. At the moment, it really appears that CHG entered southern Europe, and 50% of that in central Europe, via the Balkans."
I am not forgetting at all about the Danube. There have been multiple waves of expansion along it, including LBK, possibly Lengyel and Gatersleben. Actually, before the Steppe DNA came out last summer, my prime suspect for Indeoeuropeanisation had been the Baden-Boleraz complex.
However, we have a reasonably good time series for the Balkans, or more specifically Hungary, already available. Dave's K10 shows the following admixtures (CHG %/ EHG %/ SW_As%):
- Hungary_EN 0.71/ 1.75/ 1.00
- Hungary_CA 3.29/ 1.61/ 1.22
- Hungary_EBA 3,28/ 16.06/ 0.00 (I1502)
- Hungary_LBA 11.16/ 16.22/ 0,11 (I1504)
- Hungarian 27.69/ 13.93/ 3.64
For comparison:
- BB_Czech 24.32/ 19.31/ 0.00
- Unetice_EBA 22.42/ 24.10/ 0.65
- Czech 27.37/ 15.69/ 1.87
- Skythian_IA 28.18/ 31.50/ 3.19
- Serbian 29.90/ 11.17/ 4,85
- ItalBergamo 27.36/ 7.06/ 6.74

There is some trickle of CHG already during MN/CA, but the main CHG wave up the Danube apparently occured from the Middle Bronze Age onward. This corresponds to archeological findings indicating a spread of cremation-based practices up the Danube, which fused with the Central European Tumulus Culture into the Urnfield Culture.

Some more Hungarian aDNA incorporated into Dave's spreadsheet would help for a more fine-tuned analysis, especially as concerns the unexpected EHG boost already present during the EBA, which fades away after the LBA. Looking at our IA Skythian, I would rather have expected Hungary to increase in EHG admix during the Iron Age and by Hunnic/ Magyar incursions. Instead, we seem to have had a population coming into Central Europe that was CHG-strong, poor in EHG, but including quite some SW Asian admix.

In any case, when taking Czechs as Central European proxy, sizewise some 90% of their CHG was already present with Bell Beakers and did clearly not arrive via the Danube. The Danubian entry was late, around the Middle Bronze Age, and relatively soft. But we seem to have had a third CHG wave that came in even later, or at least didn't affect Hungary and Czechs before the Iron Age. That wave looks surprisingly East Mediterranean, and like having arrived on the Adriatic and Tyrrhenian coasts.

Yes there was certainly some GAC movement to SE, but wasn;t this reversed ?
Not that I am aware of, archeologically speaking. But we had some earlier innovation spread out of the Pontic that appears to have followed routes to the (north-)east of the Carpathians:
- Millet farming (6th M)
- "Southern" East European Ceramics, spreading from the North Pontic to the Narva Culture and ultimately Ertebolle (late 6th/ early 5th M)
- CT ceramic recipies entering (Nordic) Funnelbeaker (early 4th M)

We urgently need 6-4thM aDNA from Poland, the Baltics and West Ukraine to establish the extent these innovation flows already carried CHG into the Southern Baltic. Note the 4.56% CHG admix Dave's K10 shows for LBKT_EN (I0176), plus 7.06 in I0048, 4.41 in I0821, 4.16 in I0045. There is a good chance that half of the CHG had already been in Poland long before it is noted as part of the CW phenomenon.

Ryan said...

I'd be careful about assuming too much about Bell Beakers as a whole based on the samples we have from their eastern fringe. They look like they are ~50% Corded Ware admixed already, and may not be representative of Bell Beakers as a whole. Add western Europe to the list of places we badly need more aDNA from lol

Ryan said...

Also, @FrankN - I wouldn't look to X2 and the like as signs of any sort of trans-Atlantic movement, but rather structure in the original migration waves to the Americas. David Reich and others identified a "second" concurrent wave of migration that followed an inland route that was generally limited to the northern half of North America. http://www.nature.com/nature/journal/v488/n7411/abs/nature11258.html - see figure 2.

Remember, Ancient North Eurasians are the ancestors of both Europeans and indigenous peoples of the Americas, and haplogroups Q and R are close relations. That's the reason for the genetic similarities.

MtDNA haplogroup C4 is another one to look at. It correlates with X2, and shows up both in North America and in early kurgans.
https://www.researchgate.net/profile/Lancioni_Hovirag/publication/260104535_Mitochondrial_Haplogroup_C4c_A_Rare_Lineage_Entering_America_Through_the_Ice-Free_Corridor/links/0c96052f8f7b9c78aa000000.pdf

I haven't seen any evidence of R1b that wasn't clearly derived from Europeans - if you know of some, I'd be interested in taking a look. I wouldn't dismiss the possibility out of hand, but the fur trade actually reached the region by the late 17th century, and it had a pretty huge impact.