search this blog

Wednesday, December 9, 2015

Mixed marriages on the early Eneolithic steppe


It looks like Sredny Stog was the early vector for the spread of both Anatolian Neolithic and Caucasus hunter-gatherer (CHG) admixture onto the steppe, from the west and east, respectively:

These data testify the assumption about the existence of mixed Tripolye-Sredniy Stog marriages, because Tripolye population represented the Mediterranean anthropological type according to the not numerous Tripolye burials (Потехина 1999, c.154). It is interesting, that the massive Protoeuropoid type was typical for the oldest and the most eastern monuments of Sredniy Stog, while mesomorphic Mediterranean type was typical for the Igren cemetery, which was one of the youngest monuments related to the second and third periods of the Sredniy Stog culture and synchronous to the Tripolye B I and B I-II.

...

Appearance of pottery with pearls at the settlements of the third period of Sredniy Stog culture and glossy ceramics without ornamentation in the eastern variant sites, as well as the group of vessels with the steppe traces at the Svobodnoe settlement, allow me to assume the existence of mixed marriages between the Sredniy Stog and Northern Caucasus population.

Source: Early Eneolithic in the Pontic Steppes, book by Nadezhda Sergeenva Kotova, available at Academia.edu here.

See also...

The beast among Y-haplogroups

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

Who's your (proto) daddy Western Europeans?

204 comments:

1 – 200 of 204   Newer›   Newest»
Romulus said...

This paper is really interesting. So the Sredny Stog people traded their seashell pots to the Tripolye people for metal objects. I've only read the part on interactions between Tripolye/Stog but everything in there direct contradicts Gimbustas's "Steppe Theory". Still clinging to that David?

Davidski said...

Nothing we've seen lately contradicts the Kurgan theory of Indo-European origins.

Seems to me like you're clutching at straws to find something that isn't there.

Rob said...

I think that the north & western Black Sea areas will be more complex and diverse than the Samara Yamnaya samples. When we get aDNA from the Black Sea countries and further up in eastern Europe at good resolution, we'll be able to come up with final conclusions, fine-tunings and modifications.

Romulus said...

What is your definition of Kurgan theory? From my understanding it involves metal working & a chiefdom social structure being introduced into Europe from the Steppe by the Yamnaya culture. The role of the Kurgan is an elite burial for the chief.

According to the paper you posted metal working and a chiefdom social structure were characteristics of the Tripolye culture and were diffused in to the neighbouring Sredny Stog culture. Even the burial types of the post-Tripolye contact Sredny Stog are derivative of Tripolye. The paper states that elite burials are derivative of those in the Varna culture and that the Kurgans themselves have no special significance other than being burials for members of the Sredny Stog tribe who had to be buried away from the main tribe cemetary. https://en.wikipedia.org/wiki/Varna_culture

This paper clearly describes a situation in which Tripolye was the source of the bronze age in the Steppe. I have a hard time believing that every part of Steppe culture was derivative of Tripolye except the language.

Davidski said...

Proto-Indo-European people and languages were North Eurasians with a Caucasus substratum who expanded with the Kurgan cultural package. That's what linguistics and ancient genetics show.

Evidences of cultural appropriation from the western edges of the steppe and even the Balkans don't change this.

Rob said...

Kotova's conclusions are certainly not isolated. The formative role of CT and Majkop on the steppe has been highlighted time and again; and probably served to form a north IE culture from what might have originally non-IE one. This is what archaeology shows, and I suspect archaeogenetics will not be inconsistent with it. But ultimately, answer for the direction of language spread in the formative 4000-3000 BC period is an open one, if one appraises the evidence objectively.

Davidski said...

Archaeogenetics shows groups rich in R1a and R1b with maternal admixture from the Caucasus and western Black Sea area expanding in all directions at the right time to be the Proto-Indo-Europeans.

Any theories significantly deviating from this scenario are based on false hopes.

Rob said...

Well, its not about 'hopes' but objectivity and open mindedness. How positive are we that M417 and L51 will come specifically from the Dniester-Don open steppe region ?

In fact, the classic Kurgan theory stipulates these lines should come form the Volga - Caspian steppe. Where are they ?

ryukendo kendow said...

@ Davidski

Hi David, you mentioned:

"By the way, I think one of the big problems is that the Anatolian farmers have quite a bit of CHG, maybe over 10% on average. However, this is hidden for them within the Anatolian/LBK cluster. But it might not be hiding for present-day Europeans even if their ancestors also acquired it during the Neolithic, because they don't share as much drift with the farmers as the farmers do with each other."

Hmm, just to make sure we are using clear language with clear minds, what do you mean exactly? The Anatolian farmers contain 10% CHG compared to who? To their EEF cousins further west?

I mean, aren't the Anatolian Neolithic 100% EEF by definition? Which would mean that we would need some new ancient that is even further away from CHG compared to Anatolia Neolithic in what is currently a 'no man's land'/empty space to 'reveal' that Anatolia neolithic is CHG-shifted compared to this new sample. Or do you mean that some of the Cardial EN samples are indeed shifted away from CHG compared to Anatolia Neolithic, in addition to being WHG-shifted? Do you have any evidence to this effect?

Or do you mean that the EEF component in your recent run is too 'small', and too specific for recent shared variation in EEFs, such that it 'retreats' from modern Europeans, and the resulting excess EEF in moderns gets dumped into CHG + WHG?

(Just so the rest of you know what I mean, Dodecad had an East Euro cluster that, in addition to tracking steppe ancestry, was extremely specific to recent shared drift/ancestry in slavs, i.e. was too 'small' and specific, so it retreated from western Europeans and the resulting excess steppe ancestry was dumped into Gedrosia, the next closest thing. It also retreated from Caucasus populations and exposed 'West_European' there.)



Davidski said...

Considering that only a small percentage of Y-DNA lineages survive to the present, and only a small percentage of individuals will ever be tested, that's not an obstacle to sealing the Kurgan hypothesis, especially since we also have genome-wide DNA, with the least admixed Corded Ware coming out 70%+ Samara Yamnaya.

Davidski said...

rk,

Yeah, I'd say all of the Neolithic samples have some CHG, even those that come out 100% EEF. It's just not showing up in ADMIXTURE results, because of drift and pseudo-diploid calls. Note that Stuttgart almost always shows some CHG-related stuff, and is the only real diploid LBK genome we have.

a said...

Rob said...

" Well, its not about 'hopes' but objectivity and open mindedness..."
PIE issue aside.

The Khvalynsk/ Samara R1a/b + samples have opened Pandora's box for those doubting how ancient R1[a+b] is in this general region. I think your going to be pleasantly surprised when more results come in, as to the antiquity of the ancient R1a/b* hunter-gatherer lines in the region.
For example Khvalynsk was L754+ while Samara was positive for L278, both of these are in the range of 16k-18k+/-. Yamnaya are positive for snp's downstream ie 2103+/in the range of 6k+/-. All 2103+ and L51 plus have L278+L754 in basal postion-thats about 110 million+/- men in Europe. I'm quite certain the same pattern will emerge for R1a, as it spread from this region across vast distances both East and West. Separately this might not have been such a big deal. However together these ancient R1a/b Hunter/gatherer samples point at more upstream and downstream markers will turn up in this ancient R1 region. One day we may even be fortunate to get a actual R1* sample.
http://www.yfull.com/tree/R1a/
http://www.yfull.com/tree/R1b/

ryukendo kendow said...

@ Davidski

So you mean that different EEFs have different amounts of CHG, and so we don't really have a 100% 'EEF' genome as of now? While ADMIXTURE detects the extra CHG for Europeans, but does not find this for ancients? Sorry for being so nitpicky, I just think the definitions need to be clear so we don't confuse people.

Rob said...

@ -"a"-

The question is not about the ‘antiquity’ of R1 in eastern Europe – which is not doubtable. The question is where did the successful/ ‘lucky’ lines expand from ?
I doubt it was Samara, or even Khvalynsk, as classical models suggest.

Ryan said...

Does anyone have a link to some good sources on the origins of the Sredniy Stog culture they could share?

@a - we know the Steppe in this period was not the only home of R1b at least, as there's a V88-related sample in Iberia around the same period, and V88 in Africa is almost certainly just as old. Don't be surprised when R1b starts showing up in ancient remains from Sredny Stog's neighbours too. At least that's my bet.

I have a stupid question though - when the Black Sea was lower (say ~6000 BCE), would the Dniester and Danube have met?

@Davidski - Is there any reason why we ignore Y haplogroup G2a? It seems like a good candidate for a CHG marker wouldn't it? And it's pretty ubiquitous in the Neolithic.

Also, I think it's worth mentioning this quote from the Wiki page for asabiyyah:

"Ibn Khaldun argues that each dynasty (or civilization) has within itself the seeds of its own downfall. He explains that ruling houses tend to emerge on the peripheries of great empires and use the much stronger `asabiyya present in those areas to their advantage, in order to bring about a change in leadership. This implies that the new rulers are at first considered "barbarians" by comparison to the old ones. As they establish themselves at the center of their empire, they become increasingly lax, less coordinated, disciplined and watchful, and more concerned with maintaining their new power and lifestyle at the centre of the empire—i.e, their internal cohesion and ties to the original peripheral group, the `asabiyya, dissolves into factionalism and individualism, diminishing their capacity as a political unit. Thus, conditions are created wherein a new dynasty can emerge at the periphery of their control, grow strong, and effect a change in leadership, beginning the cycle anew."

Is it just me, or does this seem to apply to pre-historic cultures too? IE that the "next" wave of cultural/linguistic expansion seems to come from a group at the interface between two existing cultures.

Davidski said...

G2a isn't an Indo-European marker, because the Indo-Europeans made a big impact on Europe and Asia, while G2a did not. It's obviously a Neolithic farmer marker that was almost totally replaced by Indo-European lines.

It's also not a CHG marker. J appear to be a CHG marker, based on the two CHG genomes we have, while G probably expanded into the Caucasus with farmers.

Davidski said...

rk,

The Anatolian farmers certainly have some CHG admixture. That's probably how they also acquired J2a.

Stuttgart almost certainly has some CHG admixture. It always seems to show up in ADMIXTURE runs. This suggests that other LBK farmers also have it, even though it doesn't show up, perhaps for technical reasons?

I'm not sure if the Spanish farmers have any CHG. They might not if their ancestors were island hoppers and came from, say, the Levant rather than Anatolia.

Arch Hades said...

Can we be sure they had EEF ancestry without their genomes though? My guess is this will come up as the "Caucasus" component, not the Anatolian one.

I remember Dienekes posted an article from physical anthroplogy reviewed by J.P. Mallory that the "Mediterranean" types present in Sredny Stog represent geneflow from the Balkan Tripolye in that area.

http://dienekes.blogspot.com/2005/04/anthropology-of-sredny-stog-and.html

But then again I think the EEF [Anatolian] component is just craniofacially similar to the Caucasus folk so these conjectures are probably wrong IMO. We would see EEF in Yamnaya if it was in Sredny Stog.

Rob said...

Ryan

Srednij stog is used as a catch all term for the Eneolithic period (4200-3000 BC) in Ukraine; as is Khvalysnk for western Russia .
So it's a pretty broad term, made more complex by the fact that different archaeologists subclassify and periodise each component site, ceramic complex and subregions differently

But I'd recommend you look at Rassamakin - he's got the best grasp of all- and his stuff is all available on line (academia.edu)

Arch Hades said...

Proto-Indo-European people and languages were North Eurasians with a Caucasus substratum who expanded with the Kurgan cultural package. That's what linguistics and ancient genetics show.

Maikop was Kurgan too, but you think it's non IE, and proto IE was formed north on the steppe, right? So which people do you think started the Kurgan tradition?

Kapak said...

""Maikop was Kurgan too, but you think it's non IE, and proto IE was formed north on the steppe, right? So which people do you think started the Kurgan tradition?""

Obviously he's gonna go with whatever strokes his ego the hardest. I must say it's been amusing watching Davidski's metamorphisis over the years, from getting his hopes up about R1a only to be shot down by R1b folk.... to now Caucasus people coming into play when this whole time he's been talking down on them xD. Only way he can *somewhat* redeem himself is if the "PIE men take Caucasus women" scenario turns out true, because he's really counting on it.

Davidski said...

Very few linguistics and archaeologists consider Maikop Indo-European. According to Soviet archaeologists, Maikop settlements were actually raided by Kurgan people from the steppe.

And I'm still not seeing R1a-M198 as anything but Proto-Indo-European. You would really have to have your head deep up your ass to think at this stage that it wasn't a Proto-Indo-European marker.

Rob said...

Arch

There were many different types of kurgans; and it's just one component of many different cultural elements; so it's not the be all and end all

But, the specific types of kurgans seen in Yamnaya have theor origins in the CT and majkop cultures

Arch Hades said...

Actually "PIE men" had Caucasus mothers. And these later R1 PIE expanding males still had this ancestral line from their mothers but carried a lot of R1 from their fathers. As I've said before Mama genes count too. I don't think it was all EHG males and Caucasus women forming the PIEs, but the paternal/maternal contribution could have been pretty lopsided. That's something we need more samples get at least a clearer picture of. I think we need a lot more samples of CHGs and the Maikop culture, to see what Y lineages the males in that culture carried.

So it was EHG North Eurasiatic men who were taking Caucasus women to create the later PIE expanding men. But anyway, I dunno...i don't know if I should view EHGs as patriarchal macho men. I've heard they were more probably like wanderers on shamanic journeys and stuff.

Rob said...

Dave

"I'm still not seeing R1a-M198 as anything but Proto-Indo-European. "

Surely not. You do realise M198 formed 14ky BP? These fisher-forager-hunters werent PIE; I'm pretty sure. If you mean Z280 lineages spread certain PIE then one can easily agree

Arch Hades said...

Last reply of mine was to Kapak.

"But, the specific types of kurgans seen in Yamnaya have theor origins in the CT and majkop cultures."

So basically the Indo-Europeans to their cultural tradition of Kurgan burial mounds from their southern Caucasus ancestors? Is there an academic source that states this. Anyway it's pretty interesting.

Davidski said...

Sounds like bullshit to me, probably based on bloated C14 dates for Maykop.

C14 dates for Repin Kurgans on the Caspian steppe are as old as for Maykop. So why would the Yamnaya tradition come from Maykop and not Repin?

Rob said...

Arch

"
So basically the Indo-Europeans to their cultural tradition of Kurgan burial mounds from their southern Caucasus ancestors?"

Not exactly. As I said; only some kurgans appear to be of Caucasus (Majkop type); some are modelled on CT types. But who knows how this translates at an overall level? There are many other factors to consider - like economy, religion, language , etc

The point is : the proto- Yamnaya steppe was subject to massive influences in its formative period (4000-3000); before it appears to have in turn expanded back out. But the crucial point is that no linear continuity can be traced back to pre-4000 BC cultures of the steppe- which were foragers and thus not IE per definitionem

Kapak said...

"" I dunno...i don't know if I should view EHGs as patriarchal macho men. I've heard they were more probably like wanderers on shamanic journeys and stuff.""

Wanderers sounds about right. & maybe even wander on to leave his hypothetical Caucausus wifey to rear the children herself?..... and you know who spoon-feeds and teaches kids how to speak words, right? ;)

I suggest you go on google, and give John Colarusso's "Pontic" language family a whirl...... he foretold of the N-W Caucasian - PIE connection long before any Reich et en al. or what have you. Mind-blowing.... if you can keep an open mind.

Kristiina said...

In “Comparative Indo-European Linguistics”, Beekes identifies IE words with comparative method. I picked up the words that are attested both in the West and the East and that should give a hint of the character of the Indo-European culture: *h2éios ‘bronze’, *gweh3us ‘cow’, *péku ‘livestock’, *h1ekuos ‘horse’, *kuoon ‘dog’, *gwreh2uoon/*gw€rh2n ‘quern’, *h2egros ‘fallow land’, *Hroth2 ‘wheel’, *h2eks- ‘axle’, *uegh- ‘to carry, ride’, *neh2us ‘ship’, *kwrih2- ‘buy’.

I did not pick up words with a limited distribution or words that are difficult to reconstruct due to irregularities as this could point to the word being a loanword.

To me the IE culture looks like being very mobile (wheel, axle, verb ‘ride’, ship), and IE’s kept animals and raised cattle (dog, cow, livestock, horse, fallow land) and were involved in trade (buy). Agricultural terminology is not shared to a great extent as I could find only the word for quern.

Rob said...

The aforementioned "palaeolexemes' suggest that PiE spread after the secondary products revolution, which occured everywhere; not just the steppe. In fact, the steppe *received* these inputs, and did not invent them

From where ? Two routes

Davidski said...

I thought that this pointless idea of trying to put the Proto-Indo-European homeland south of Europe and the origin of R1a in India was going to end after a whole bunch of Eastern European hunter-gatherers and Kurgan chiefs were shown to carry R1a and genetic profiles closest to present-day North/East Europeans.

Nope. It looks like it'll go on for a while yet.

Balaji said...

Ryukendo Kendow and Davidski,

Thank you for the very interesting exploration of Paniya and Austroasiatic using D-statistics that you did a little while ago. I thought that perhaps we could revisit this.

Dai Paniya Kotias Chimp -0.0132 -3.247 113679
Dai Paniya LBK_EN Chimp -0.0092 -2.997 134550
Chimp Ust-Ishim Paniya Loschbour 0.0021 0.315 113731
Chimp Ust-Ishim Austroasiatic Loschbour 0.0001 0.009 112322
Chimp Ust-Ishim Paniya Kostenki14_UP 0.0041 0.549 124370
Chimp Ust-Ishim Austroasiatic Kostenki14_UP 0.0032 0.423 122830

The above statistics indicate that Paniya and Austroasiatic have some ANI (West Eurasian) ancestry. Then there are the following numbers.

Chimp Ust-Ishim Paniya Loschbour 0.0021 0.315 113731
Chimp Ust-Ishim Austroasiatic Loschbour 0.0001 0.009 112322
Chimp Ust-Ishim Paniya Kostenki14_UP 0.0041 0.549 124370
Chimp Ust-Ishim Austroasiatic Kostenki14_UP 0.0032 0.423 122830

Kostenki14 Kotias Ust-Ishim Chimp 0.0273 3.758 373910
Loschbour Kotias Ust-Ishim Chimp 0.0305 4.665 404058
Chimp Ust-Ishim LBK_EN Loschbour 0.0151 2.692 279178
Chimp Ust-Ishim Kotias Loschbour 0.0305 4.665 404058

The above statistics show that the ANI in Paniya and Austroasiatic is different from both Kotias and LBK_EN. This was also shown by Davidski's Smarter Bear Plot with Satsurblia and Kotias.

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

According to Moorjani et al. Paniya have only 17% ANI. The ANI in Austroasiatic is probably even less. It will be good to see if how Ust-Ishim chooses between populations with higher and perhaps different kinds of ANI. Could we got the following D statistics?

Chimp Usi-Ishim Pulliyar Loschbour
Chimp Usi-Ishim Pulliyar Loschbour
Chimp Usi-Ishim Chamar Loschbour
Chimp Usi-Ishim Kanjar Loschbour
Chimp Usi-Ishim Kol Loschbour
Chimp Usi-Ishim Indian_Singapore Loschbour
Chimp Usi-Ishim South_Indian Loschbour
Chimp Usi-Ishim Kusunda Loschbour
Chimp Usi-Ishim Bengali_Bangladesh Loschbour
Chimp Usi-Ishim Burusho Loschbour
Chimp Usi-Ishim Brahmin_UP Loschbour
Chimp Usi-Ishim Brahmin_TN Loschbour
Chimp Ust-Ishim Brahui Loschbour
Chimp Usi-Ishim GujaratiA Loschbour
Chimp Usi-Ishim Kalash Loschbour
Chimp Usi-Ishim Makrani Loschbour
Chimp Usi-Ishim Sindhi Loschbour

Thanks.

Colin Welling said...

@Romulus

I have a hard time believing that every part of Steppe culture was derivative of Tripolye except the language

Linguists pretty soundly place PIE in the steppe. Knowing that metal came from CT culture changes nothing.

What do you mean every part of the Steppe culture was a derivative of Tripolye? Ive never heard a credible archeologist ever say anything like that. Gumbitas could be totally wrong but that doesn't change the basic kurgan hypothesis that PIE was from a steppe culture they obviously spread that culture all over the place.

Gumbitas was fucking nuts, and her version of the Kurgan theory isn't important IMO.

The DNA is showing us exactly what Davidski has said. The heritage that ties the yamnaya together and defines the majority of their ancestry is EHG and CHG. The mix of EHG and CHG in yamnaya mirrors the northern (uralic) and southern (caucasian) linguistic border that was predicted by the steppe hypothesis. And for what its worth, EEF is not spread across the steppe at the time of the yamnaya, so its not like you can use a demic model to say that yamnaya got their IE from CT.

Im guessing the western yamnaya had a chunk of EEF but EEF and CT isn't the dominant influence that unites the yamnaya.

Davidski said...

The CHG K8 is now the CHG K10. Scroll down to the update here...

http://eurogenes.blogspot.com.au/2015/12/sardinians-pack-lot-of-chg-ancestry.html

Rob said...

@ Arch

"I've heard they were more probably like wanderers on shamanic journeys and stuFF"

Heard where ? Yamnaya were specialist pastoralists, mobile, and apparently rather healthy.


@ Colin

"Linguists pretty soundly place PIE in the steppe."

More soberly, the linguistic evidence points to somewhere in northwest Eurasia; during the Copper age. There is not demanding it be specifically the steppe

*"Knowing that metal came from CT culture changes nothing"

(I know this wasn't your claim). The metal situation is rather complex. Briefly, pre -4200 BC, steppe communities got metal from the Balkans. This shifts post-3800 BC; when north Caucasian forms form a monopoly over much of north Eurasia - from the Danube to the Urals. INterestingly, in Corded Ware regions types reminiscent of the older Carpatho-Balkan forms appear to have continued

Ryan said...

@David - I think you mistake me on G2a. I'm not suggesting G2a is the only CHG marker, nor am I saying it's a particularly relevant one for the Indo-European expansion. We only have a sample size of 2 for CHG. I may be pretty far off base for G2a though - let me think about it more. It just kind of made sense in my head - G is very "basal Eurasian" in a Y-DNA tree sense, and it peaks in the Caucasus... Anyways, let me think more.

I think we can at least assume H13 as a whole is a CHG marker? I'd note that it has a Mediterranean coastal distribution in Europe, and higher frequencies in Sardinia (at least according to Eupedia - take that as you will).

So CHG women got to Sardinia by boat?

Ryan said...

Also, if Sredny Stog were mixing with CT people, why is there 0 EEF in Yamnaya? Shouldn't there be some?

Looking at your latest K10 run - I'm having a hard time making sense of it. Based on the FST distances, Anatolian Neolithic, SW Asian and CHG are somewhat closely related to one another. WHG is also closer to SW Asian and CHG than to Anatolian Neolithic. That doesn't make much sense does it? Or does it simply imply greater African or some other "basal" admixture into Anatolian Neolithic? Though the FST distances apply otherwise.

Also, if we go from the Iberian Chalcolithic to the Basque, we see a huge increase in CHG with a comparatively modest increase in EHG still. I can buy that to an extent for the Basque - presumably they got a small amount of EHG and CHG together from mixing with Indoeuropeans, and an excess in CHG from whatever delivered CHG to Sardinia, but there are a lot of Indoeuropean speaking groups with large CHG values and virtually 0 EHG. I think the different clusters are bleeding over into one another somehow. Maybe Treemix would be more suitable as a primary tool here?

Like, look at the Corded Ware samples, or even Bronze Age Hungary - there's relative parity between CHG and EHG, just like with Yamnaya. Now look at the modern populations in Central/Western Europe - they generally have a lot more CHG than EHG. So there's some massive CHG expansion after the expansion of Indoeuropean?

Maybe the EHG samples we have aren't very good proxies for the actual EHG group that led to Proto-Indoeuropeans? Karitiana is coming out as 40% EHG, and with Q1a showing up in EHG, maybe these Karelians were a bit more closely related to Amerindian groups than the EHG on the Pontic Steppe. I dunno, confusing to say the least. Thoughts?

Davidski said...

Ryan,

G2a just looks like an early farmer lineage to me that also penetrated the Caucasus. The reason I say J is a CHG marker is because we have 2/2 CHG samples belonging to J, and also that CHG and J reach high frequencies in the Caucasus today.

There is J in Sardinia, so both CHG males and females made it down there.

Balaji,

Chimp Ust_Ishim Pulliyar Loschbour 0.001 0.145 112449
Chimp Ust_Ishim Pulliyar Loschbour 0.001 0.145 112449
Chimp Ust_Ishim Chamar Loschbour 0.0019 0.314 112450
Chimp Ust_Ishim Kanjar Loschbour 0.006 0.951 112450
Chimp Ust_Ishim Kol Loschbour 0.0042 0.672 112450
Chimp Ust_Ishim Indian_Singapore Loschbour 0.0134 2.43 368472
Chimp Ust_Ishim Kusunda Loschbour 0.0064 1.132 501362
Chimp Ust_Ishim Bengali_Bangladesh Loschbour 0.0069 1.227 501362
Chimp Ust_Ishim Burusho Loschbour 0.0109 2.06 501362
Chimp Ust_Ishim Brahmin_UP Loschbour 0.0067 1.101 112450
Chimp Ust_Ishim Brahmin_TN Loschbour 0.0086 1.267 112418
Chimp Ust_Ishim Brahui Loschbour 0.0211 3.903 501362
Chimp Ust_Ishim GujaratiA Loschbour 0.0135 2.317 501362
Chimp Ust_Ishim Kalash Loschbour 0.015 2.661 501362
Chimp Ust_Ishim Makrani Loschbour 0.0272 5.041 501362
Chimp Ust_Ishim Sindhi Loschbour 0.0136 2.563 501362



Kristiina said...

Where and when do we we have the first archaeological evidence for wheel and wagon? It looks like wheel and wagon were important for the expansion of IE languages. However, it cannot be the whole story as wagons are quite useless in a forest, and IE languages spread for example to Scandinavia where wagons were not very practical.

How about the cow? Did Neolithic Europeans keep cows?

FrankN said...

@Kristina:
- *kuon ‘dog’ is a Paleolithic Wanderwort, reflected e.g. in East Chadic *kany, Korean *kàŋ-, Nahuatl (Aztec) Itzquintli, Lakota sunka (c.f. Sanskr. sunaka), and various more (see link, root *KVNV).
http://starling.rinet.ru/cgi-bin/response.cgi?baseid=+0+1+2+3+4+5+6+7+8+9+10+11+12+13+14+15+16+17+18+19+20+21+22+23+24+25+26+27+28+29+30+31+32+33+34+35+36+37+38+39+40+41+42+43+44+45+46+47+48+49+50+51+52+53+54+55+56+57+58+59+60+61+62+63+64+65+66+67+68+69+70+71+72+73+74+75+76+77+78+79+80+81+82+83+84+85+86+87+88+89+90+91+92+93+94+95+96+97+98+99+100+101+102+103+104+105+106+107+108+109+110+111+112+113+114+115+116+117+118+119+120+121+122+123+124+125&root=config&morpho=0&off=&text_proto=&method_proto=substring&ic_proto=on&text_meaning=dog&method_meaning=substring&ic_meaning=on&text_any=&method_any=substring&sort=proto&ic_any=on&first=1
Actually, I see it as a composite of *Ku, which means "to move, run, depart" in various language families (and is even more frequent as part of 'dog terms' around the world), and Proto-Uralic *ni (or related vowel) that shows up in various roots related to "fur, skin, de-skinning". Thus, PIE *kuon "dog", as Uralic *kujna "wolf", would both mean "fur-bearing runner (hunter)".
[I have written a bit about Dog aDNA at the end of the post on the Jäger paper, you might want to consult this as reference].

- The "wagon terminology" (including wheel/ axle) has been discussed ad infinitum, no need to repeat it here. Lets just note that the *uegh- ‘to carry, ride’ concept seems to be most stringently applied in Germanic (cf. way, wagon, to wage, wave, weigh - the latter seems to point at the first wagons only beeing two-wheeled), which corresponds to the first evidence for wheeled (ox-)carts stemming from Funnelbeaker (FB) contexts. However, unless FB spoke PIE (which I don't think, as Germanic languages are lexically, morphologically and phonologically quite removed from the IE 'mainstream'), this "PIE terminology" would actually constitute an early borrowing from a Non-IE language.

- *gweh3us ‘cow’, *péku ‘livestock’, *h1ekuos ‘horse’, *kwrih2- ‘buy’ are semantically linked. The "cattle trade" connection is well documented, be it for Latin pecus "cattle" vs. pecunia "money", or for Ags. cheap "cows, cattle" turning into English cheap (i.e. 'good business'). "buy" vs. Celtic bou "cattle" is also pretty transparent in this respect. This doesn't invalidate the IE connection (and certainly not your conclusion on cattle herding at the heart of IE culture), but leaves us with a couple less of truly independent cognates.

- The Bouckaert paper on Northwest-Caucasian - IE connection that Dave posted here some months ago didn't convince me at all when it comes to lexical parallels identified therein. There, were, however, two exceptions: One was kinship terms (relevant to Dave's opener!), the other one was animal husbandry terminology. IIRC, *h2egros ‘fallow land’ was part of these NWC-PIE cognates. Others which I recall were 'porch', 'meadow', and 'ewe'/'ovis'. Note on the latter the archaic Caucasian nominal class system, often fossilised but still reasonably well preserved in Chechen. There, a 'null' or 'e-' feminine prefix contrasts with a 's-' masculine prefix. Technically, this would allow to turn a female 'ewe' into a male 's-ewe', i.e. a 'sheep' (which originally designed the male, before beeing gender-neutralised). The same principle would turn a German Tier "deer, animal" into a Stier "bull" (Stu in Chechen).

What does that leave us with: PIE as quite some Creole, drawing its 'core vocabulary' from various sources.

ryukendo kendow said...

@ Davidski

David, let's clarify the terms a bit; let's set 'CHG' to refer to whatever is more related to Kotias and Satsurblia than anything else, 'EEF' to refer to whatever the Neolithic Farmers of Europe shared with each other to the exclusion of all other genomes including CHG.

By this definition, whichever Neolithic farmer is the furthest from CHG automatically becomes the sample around which EEF is defined. If we discover some sample from the levant/cyprus, and this sample is further away from CHG than all other Neolithics, then this automatically beccomes the new 'EEF', and all other farmer genomes are 'revealed' to be mixtures of this 'new' EEF plus CHG.

I agree that, going even by our current definitions, Central European neolithics seem like they are a bit CHG, while Cardial Neolithics seem a bit 'purer'. Stuttgart scores in the 'Caucasus' component in other runs as well, in constrast to other neolithics.

So much for EEF and CHG. OK, now looking at the previous stats. It seems that Armenian are quite Mota-shifted compared to Georgian, so they may indeed have some African ancestry from the 'SW Asian' component, or whatever it is that differentiates them from Georgians.

Since you seem to have lists of populations sorted by Geography, would you mind repeating the statistics with a list of Middle Eastern populations in the Dataset? Statistics of the form:

Primate_Gorilla Mota X LBK_EN
Primate_Gorilla Mota X Kotias
Primate_Gorilla Mota X Adygei
Primate Gorilla Mota X Austroasiatic_Ho
Primate_Gorilla Mota X Dai

where X is the Middle Eastern population.

David, it occurs to me that XY plots of
Primate_Gorilla X Armenia_BA Chimp vs
Primate_Gorilla X Armenia_IA Chimp

as well as

Primate_Gorilla X Armenia_BA Chimp vs
Primate_Gorilla X Armenian Chimp

Where X is all populations, would also be very interesting.

Now for your newest ADMIXTURE run, which is very interesting. Excellent David!
I still strongly suspect that the EEF and EHG clusters are too 'small' and the WHG and CHG clusters are too 'large' in both your runs, because EEF and LNBA are set apart by their large number of ancients in which to dominate, while WHG and CHG are distinguished by only 2-3 ancients each. I think this causes the 'CHG' and the 'WHG' clusters to overspill into moderns, while EHG and EEF 'retreats' from moderns into ancients. Why not do a supervised run, where you get to determine what the components are? I wonder how that would turn out.

On the other hand, the 'Afro-Eurasian' component that peaks in Hadza and is well represented in Africa is very interesting, especially the fact that almost all the Sub-Saharan in the Middle East is revealed to be 'Afro-Eurasian'. E.g the Egyptians have almost no 'Sub-Saharan', but do have ~15% 'Afro-Eurasian', and this is extremely evenly spread out, and so cannot be a signal from the slave trade. Likewise in Palestinians and so on. Very interesting that the Austroasiatic_Ho score ~5% in 'Afro-Eurasian', and also non-zero 'SW Asian' whatever that means. I guess the Basal in South Indians is indeed weird.

David, is it possible to run a 'components of components' test here? I.e. use some components as references, and one as a sample, one after another. What happens to 'SW Asian', 'Afro-Eurasian' etc. will be very interesting to see.

@ Tobus

I saw from the previous post that you have incorporated the new genomes into your dataset already. Cheers!

Tobus, can you perform Treemix?

Rob said...

Kristiina

Wheels were probably invented in Mesopotamia (kohl) but spread rapidly and widely through a cultural transmission. They were certainly present in Cucteni-Tripolye and Majkop, from where Yamnaya adapted them. Each EEF group had different economies, but cattle were already present in Carpathian late Neolithic groups

Frank N

PIE is not a 'creole'. Linguists have very specific definitions of a creole, and one cannot simply make up their own theories. IE has none of the features of creoles, such as lacking inflexional morphology. However, some of the pre-PIE steppe languages might have been creoles

Kristiina said...

Frank, you may very well be right in what you say about the origin of the IE words mentioned above, but, in any case, they are shared between both Eastern and Western IE languages and show regular sound correspondences and, therefore, must have been part of the IE expansion.

Rob, do we have DNA from Carpathian Neolithic groups who kept cattle?

It is true that cattle was domesticated pretty early: "Archeozoological and genetic data indicate that cattle were first domesticated from wild aurochs (Bos primigenius) approximately 10,500 years ago. There were two major areas of domestication; one in the Middle East/Europe (the area that is now Turkey) giving rise to the taurine line and a second in the Indian subcontinent (the area that is now Pakistan) resulting in the indicine line. Modern mitochondrial DNA variation indicates the taurine line may have arisen from as few as 80 aurochs tamed in the upper reaches of Mesopotamia near the villages of Çayönü Tepesi in southeastern Turkey and Dja'de el-Mughara in northern Iraq".

Chris Davies said...

@ FrankN -
"*kuon ‘dog’ is a Paleolithic Wanderwort, reflected e.g. in East Chadic *kany, Korean *kàŋ-, Nahuatl (Aztec) Itzquintli, Lakota sunka (c.f. Sanskr. sunaka), and various more"

nkịta - Igbo

kutiru - Fula

ngui, ngitĩ - Kikuyu

FrankN said...

@Rob: "Wheels were probably invented in Mesopotamia (kohl)" That's one theory, so far unproven. Below you find a summary on all securely and tentatively dated archeological evidence for wheeled carts:
http://antiquity.ac.uk/projgall/mischka329/mischka329_supplement.pdf

Extracts:
- Wheel traces, Flintbek/D (FB), 3423–3390 calBC
- Pictogram on pottery, Bronocice/PL (FB), 3631–3380 calBC
- Wooden wheel with axle, Stare gmajne/SLO (Baden), 3357–3116 calBC (axle, old wood effect?)
- Clay tablet, Uruk, 3641–3381 calBC (unsecure: dating of cultural layer, not in situ, charcoal may display old wood effects)
- Wheel model, Ğebel Arūda/IRQ (late Uruk), 3335–3103 (charcoal, may display old wood effects, from burnt roofing)
- Wheels, Starokorsunskaya/RUS (Majkop), undated, context estimated at >3370 BC
- Terracotta-fragment of wagon model (?), Harrapa/IND, undated, context estimated at 4990 BC

This means so far we can be sure of FB having carts by around 3.400 BC. Maykop and Uruk appear to be at best contemporary, possibly later than FB. As concerns Majkop, there have been quite a number of older Russian "context datings" that had to be substantially corrected downwards over the last years.
Harappa is interesting, though, and might deserve a closer look.

Independent invention of the wheeled cart cannot be excluded. The following technological predecessors are being considered:
1. The potter's wheel (assumed for Uruk)
2. Flywheels, especially those mounted on drills (FB?)
3. Spindles (Harappa?)

@Chris Davies: When allowing for a centum-satem sound shift, the Mbutis' Ba-senji, clearly one of the most ancient dog races, may also belong to the class of 'kuon Wanderwords.
Another etymology, however, links the Basenji to Suahili mbwa shenzi “wild dog”. In any case, the Mbutis' dog keeping tradition, outside the habitat of any wolves, should give anybody using them as 'unadmixed' outgroup for admixture stats a second thought..
https://en.wikipedia.org/wiki/Basenji

Your examples seem to relate to what I think are derivatives on a hypothetical *ku "to run", as e.g. in proto-AfrAs *gay-"to move", Maa a-kúɛt "to move fast, run", and of course Lat. currere "to run" plus other IE reflexes such as Celtic carrus and Germanic hart and horse. Note here also the Kudu, Sidamo goda "game, gazelle", Akkadian gadu "(young) he-goat", and lat. haedus, engl. goat, [pol. koza?], etc. The chain can be set forth into Dravidian and Austroasiatic.

Dmytro said...

Note BTW that Kotova's work concerns the period ca. 5300-4100 BCE, long before Maykop, Repin, never mind classical Yamna...

And her ideas about the genesis of Khvalynsk are also quite significant. Cf. the Conclusion, pp. 142 ff. and the prior statements on pp. 121-123.

Now for the aDNA evidence... I suspect that both R1a and R1b buffs will be satisfied...

Romulus said...

@David

"G2a is not an Indo-European marker"

There is no such thing as an "Indo-european" marker. Indo-European is a language family and G2a is a y chromosome. They can be spread and move independently of each other. The idea of an unbreakable connection between the two is obviously false, there are many examples (Basque, Chadic). This type of thinking I see as unproductive. I could see R1b as a majority CHG marker in Europe coming from Varna, even though it exists on the Neolithic Steppe. There is a lot of it south of the Steppe that needs to be explained. I see Indo-European language and culture coming from Varna/Tripolye regardless of Y or autosomal, whether it originated there or with Farming in general I don't know. Sredny Stog types sem only a proxy for this culture swept up in the wave.

@Colin

"linguists have traced PIE to the Steppe"

Linguists speculate and not much else when it comes to PIE, they can't "prove" shit.

capra internetensis said...

@FrankN:

It does not matter if PIE borrowed the word from some other language - borrowing words for useful inventions is utterly normal and does not remotely make a language a creole - the point is whether the word shows regular development in daughter languages after being borrowed into the proto-language, rather than being borrowed piecemeal into different branches from different sources at a later date.

If for instance the root for "wagon, drive" was borrowed from Funnelbeaker - not that we have anything like the chronological resolution necessary to say where wheeled vehicles spread from - it must have been borrowed at an early stage, since we have Sanskrit vahati. It is not plausible that Indo-Aryan borrowed a Harappan word and Germanic borrowed a Funnelbeaker word and they just happen to look like regular cognates; the word is Proto-(Nuclear-)Indo-European. Which is the point, if we are trying to reconstruct PIE material culture, it doesn't matter where the words and useful objects may have come from originally.

Romulus said...

The 'proposed' linguistic connection between PIE and the Steppe is itself very stupid and has been debunked. Wheeled vehicles weren't invented on the Steppe, they were invented in Mesopotamia and transmitted TO the Steppe BY Tripolye, just like every other aspect of their culture. David Anthony's book is expired crap.

Der Frost said...

Differences between anthropological types of males (robust UP) and females (gracile Mediterranean) belonging to Late Chalcolithic / Early Bronze Age cultures of Romania-Moldova-Ukraine were noted even during the interwar period by Romanian bioanthropologist Olga Necrasova (born in 1910 in St Petersburg, emigrated with her White Russian family to Romania after WW1).

With respect to how exactly these contacts between settled and pastoral populations took place, this article is also worth a read: http://cucuteniar.ro/downloads/Dan%20Monah%20-%20Salt%20Springs%20Places%20for%20Salt%20Recrystallization%20and%20Ritual%20Centres%20for%20Exchange%20with%20Steppe%20Population.pdf

Last, to nitpick a bit, the use of the term "Tripolye" alone is too ambiguous in this context. Cucuteni and Tripolye (or Tripillia in Ukr.) were kin but distinct cultures (the Cucuteni is the originator, an expression of the Balkan Chalcolithic world and thus more complex, while Tripolye is the a acculturated form, a more rustic "forest Chalcolithic" culture). So are you referring only to Tripolye or to C-T as a whole?

Ryan said...

@David - Surely you mean J2a, and not J2a as a whole?

Also, just because CHG samples are 2/2 for J2a doesn't mean that with 100 samples of CHG would all be J2a.

Matt said...

New K10 run through the same FST distance based PCA process

PC1 and PC2 - http://i.imgur.com/xxqEOvh.png
PC1 and PC3 - http://i.imgur.com/Wsx7Lgh.png

Kurti said...

David said

"Very few linguistics and archaeologists consider Maikop Indo-European. According to Soviet archaeologists, Maikop settlements were actually raided by Kurgan people from the steppe."

Well thats what SOME Russian archeologists thought back than.

Most modern archeologists date the Maikop Kurgans as older as the Yamna Kurgans. In fact the oldest Kurgans are said to be in Leyla Tepe in Southeast Caucasus.

"The earliest known kurgans are dated to the 4th millennium BC in the Caucasus.[3]

https://en.wikipedia.org/wiki/Leyla-Tepe_culture

Obviously Kurgans are a Southern introduction into the Steppes not vica versa.

I am also not aware of any scientists having proved that all of the ~50% Caucasus DNA came via female lineages. How? If ~50% of Yamna mtDNA is actually EHG.

We are dealing here with a fusion of EHG and CHG like people. Imo the only reason we only encountered R1b in Yamna period Kurgans is because Kurgans represent Elite burials. They are not representative of the General population. Most likely those are the lineages of the "Royal families".

A Strong indication for this is how one single lineage ( R1b L23) is being replaced entirely by another single lineage (R1a z93). One Elite replacing the other. No sign of other modern Indo European lineages despit ~80% of the Indo European language family being outside of these two lineages.


To be fair we still don't even know where R1b L23 (typical for Armenian, Anatolian, West Iranic and Balkan branches of Indo European) came from.

Kurti said...

By the way, to say ~4000 BC Maykop culture was all yDNA J because 13-8.000 BC CHG samples turned out this way is just .... amateurish.

Krefter said...

@Kurti,

50% of Steppe mtDNA is not EHG. 72%(36 of 50) of Yamnaya/Poltavka mtDNA is not U(xK). Catacomb has a strangely high amount of U4 and we don't have their autosomal DNA but besides that Steppe mtDNA looks mostly CHG. Besides we can't determine if admixture was gender bias because Yamnaya Y DNA comes from a single Mesolithic man.

Kurti said...

Fire Head said

"50% of Steppe mtDNA is not EHG. 72%(36 of 50) of Yamnaya/Poltavka mtDNA is not U(xK). "


Well than it's 30%. I don't know how much this changes the point. 50 or 70. It's still 30% too much EHG female mtDNA that remains.

"Besides we can't determine if admixture was gender bias because Yamnaya Y DNA comes from a single Mesolithic man."

As I said ROYAL Burials, therefore not representative of the general population. But how do you know yDNA came from a single Mesolithic man? We are in the Bronze AGE era so it is a single Bronze Age male or rather family that probably spred it's yDNA.

Romulus said...

@Kurti

Not all kurgan burials are elite burials, likewise not all elite burials are kurgan burials.

"There was a combination of ground and kurgan burials in the funeral rite in the Sredniy Stog culture. A kurgan mound is often considered as an indication of high social status of the deceased, because it demanded considerable temporal and working expenditures. But it is difficult to explain the existence of kurgan mounds in the Eneolithic only on the basis of social status of the deceased. The burials with numerous grave goods and well-decorated funeral clothes, as well as the burials with poor inventory and without adornments, were found in the kurgans and ground cemeteries. Simultaneous use of ground and kurgan burials can’t be also explained by the gradual transition from the burials in the large cemeteries to the individual burials under kurgans, because ground burials were widely used during all period of the Eneolithic, Bronze Age and following times. In my opinion, the main explanation of the existence of such biritualism in the rite is the territorial affixment of the burial. The majority of ground cemeteries was situated near the settlements, such as Igren, Petro-Svistunovo and others. They are also encountered in the steppe between the Don and Dnieper - on the native territory of Sredniy Stog population. Small mounds above the burials near the settlements assisted in orientation in the planigraphy of necropolis."

Kurti said...

Romulus

""There was a combination of ground and kurgan burials in the funeral rite in the Sredniy Stog culture. A kurgan mound is often considered as an indication of high social status of the deceased,"

Well I should have clarified myself a bit. In my previous post I wrote "Elite". With "Royals" I meant an Elite group or as You said "High social status". We know in most ancient cultures, high social status is often shared between a few related groups/families. So having Kurgan Burials predominantly belonging to one and the same branch of the same Haplogroup is a clear indication for me that we are dealing here with Elite groups related to each other paternally.

Davidski said...

Have you actually noticed that all of the reliably tested Andronovo, Corded Ware, Poltavka, Potapovka, Scythian, Sintashta, Srubnaya and Yamnaya remains are R1a and/or R1b.

Even if these are all from elite burials, and the lower crusts of society carried more varied haplogroups, it's clear that the early Kurgan expansions were driven by the paternal descendants of Eastern European hunter-gatherers. So you have no argument.

Ryan said...

@David - I'm having a hard time reconciling those groups as ancestors of all Indo-Europeans with your K10 showing Greeks and Armenians as 0.01% EHG, Iranians as 4% EHG, and the French with 2.5 times as much CHG as EHG though. Thoughts?

Davidski said...

These people simply don't have very much genome-wide ancestry from the earliest Indo-Europeans.

That's because languages aren't passed on through genes. They are learned.

But if you want to link the Proto-Indo-European expansion with something other than the Kurgan expansion, then I'm all ears. Thing is, I haven't seen anyone put forth a sensible alternative yet, and every time we get more DNA from ancient people who are supposed to be early Indo-Europeans, the same patterns show up.

Davidski said...

rk,

Here are those D-stats.

https://drive.google.com/file/d/0B9o3EYTdM8lQYkZnQU9HNUE2M3M/view?usp=sharing

Btw, the latest ADMIXTURE run was supervised. The thing with supervised tests that use a lot of samples and clusters, is that they're not really supervised. It's more like an exercise in trying to point the software in the right direction. Also, most of the clusters aren't independent of one another, so if I try to fiddle with the levels of one, this can have a dramatic impact on some of the other clusters, or even the whole run.

I'm going to try some unsupervised runs with my synthetic CHG, EHG and WHG samples, which actually look pretty close to what they should be.

Romulus said...

My current theory is this: Indigenous peoples of Europe achieved behavioral modernity through contact with the cultures of Neolithic Farmers migrating from the Near East. These cultures served to deliver advanced Mesopotamian ideas to Indigenous Europeans.

1.Agriculture
2.Sedentary living
3.Animal domestication
4.Metal Working
5.Social Structure.
6.And because it makes sense, Language.

When Indigenous Europeans met these people they probably thought they were Gods because they were much more advanced, hence they became subservient to them and adopted their language and culture. Once the initial culture shock regressed and these ideas became fully integrated in to Indigenous European cultures, bigger Hunter Gatherer types killed off the original farmer males and took over. During this period there is an East to West migration of R1b types to Western Europe via BBC, originating in Ct/Stog. I like this theory because it is linear and very basic/simple. It does not involve the spontaneous appearance of ideas in places where there was no catalyst for them to exist. It can explain Basque as a pre-farming language.

Davidski said...

Indo-European is a North Eurasian language group, and its initial expansion is tied intimately to the expansion of indigenous Eastern European DNA across and out of Europe, especially paternal markers.

So no, Indo-European languages did not arrive in Europe with farming.

Actually, that would be quite hilarious, considering that practically all Indo-European languages contain substrata usually related to farming that is clearly of non-Indo-European origin.

What it would mean is that the supposedly advanced Indo-European farmers borrowed most of their farming related words from indigenous European hunter-gatherers.

Duh.

Rob said...

Romulus

"My current theory is this: Indigenous peoples of Europe achieved behavioral modernity through contact with the cultures of Neolithic Farmers migrating"

We should attempt to refrain from pulling theories out of our backsides :)

Karl_K said...

Good stuff Romulus. It all just makes sense.

Matt said...

Leaving aside the other sketchy stuff, Mathieson's stuff seems to have found that the SHG and WHGs and Early Neolithic farmers seem to have been about the same height (and didn't seem to have a BMI difference).

http://4.bp.blogspot.com/-asxeNB6Qwnc/VQTJYe198CI/AAAAAAAAKAo/H8VUnjYX9VU/s1600/height.jpg - for the graphic from the early version of their paper

http://www.unzcloud.com/wp-content/uploads/2015/10/Screenshot-10262015-052241-PM.png - for a newer version including AN instead.

In Mathieson's data Early-Middle Neolithic Central European Farmers seem to have been between than modern day Spanish and modern day British, so probably in the range of 173-175cm for males (genetic potential in the same environment as moderns). SHG and WHG seem around the same in the same data.

Yamnaya seem to be specifically taller. Possibly EHG specifically had a large body size, but they don't seem to have checked that out. People who seem to have a fair bit of extra EHG ancestry independently of Yamnaya, like the Mari (or Saami?), don't seem very large, on the whole.

(So what does "massive Protoeuropoid type" mean here? Large body / large face?).

Davidski said...

Matt, Maris and even Saami have a lot of East Eurasian admixture. Saami also seem to be adapted to semi-Arctic conditions.

The term massive in this context refers to the face, but apparently the Sredny Stog people were also very tall for their period.

Matt said...

Re: Sredny Stog, could mean an adaptation leading to the Yamnaya had already started by that time? Probably know more once they've sequenced a large set of EHG.

Ryan said...

@David - not doubting the Steppe hypothesis. I think it's core is pretty much case closed now. What's in doubt (in my view) is exactly how far we can take it, it's specifics, details, etc. And I was just curious as to why those weird results were there. If it's just a work in progress then yah, fair enough.

"These people simply don't have very much genome-wide ancestry from the earliest Indo-Europeans."

If that's the case and not just an artifact of the run, then these samples aren't the Proto-Indo-Europeans, and the Karelians aren't the ancestors of Proto-Indo-Europeans. They may be closely related sister groups though (Para-Proto-Indo-European?).

If that's the case, it actually would tell us something very interesting - PIE didn't just expand from the Pontic Steppe, but from a very small corner of it somewhere. Not all of the Steppe fully participated in this expansion.

Romulus said...

@Matt

I wouldn't place too much stock in the height estimates from Matheison. The GWAS height study still only claims to be able to predict 20% of human height variation and this includes SNPs studied from all human populations (I think the most significant SNP is found predominantly in Han Chinese and there were only 5 or so significant SNPs found). Applying these SNPs to WHG/SHG types who only contributed 20% or so of their genome to modern European pops isn't going to give anywhere near an accurate reading. Multiple Cro-Magnons were directly measured to be around 6"4 , and their physical type has been recorded in North European neolithic cultures (Funnelbeaker/CWC). Loschbour and Bichon were short (5"6) but you never know if shortness is a result of malnutrition or short genes, for tallness it can only be the result of tall genes. I can't remember where I read it but I'm pretty sure the Yamnaya skeletons were measured around 5"9-5"10, thinking of the guy with the club. Otzi I believe is 5"1.

Either way it can be confidently said that WHG/SHG/EHG/Yamna types were much taller than EEFs. There is a direct correlation for EEF autosomal in Europe and short stature, Sardinians are fairly definitely one of the shortest.

You can read a good description of Proto-Europoid here:

http://dienekes.awardspace.com/articles/metrictypes/

Krefter said...

@Romulus,
"When Indigenous Europeans met these people they probably thought they were Gods because they were much more advanced, hence they became subservient to them and adopted their language and culture."

I doubt that. The difference between Mesolithic Europeans and Neolithic Ageans wasn't as great as between for example, various native people and Europeans in the 1500-1900s.

There's no rule that says ancient West Asia has to be the most advanced. Since the Middle Ages that has not been the case. Everything doesn't have to come from West Asia. No one knew the other existed in ancient times. People lived ignorant of history and geography.

There's no reason why everything associated with civilization has to come from West Asia or Europeans were subordinate to West Asians. Does the fact Irish play the banjo make them subordinate to West Africans? No.


"My current theory is this: Indigenous peoples of Europe achieved behavioral modernity through contact with the cultures of Neolithic Farmers migrating from the Near East. These cultures served to deliver advanced Mesopotamian ideas to Indigenous Europeans."

The earliest evidence of metal working comes from the Balkans not West Asia.

" Once the initial culture shock regressed and these ideas became fully integrated in to Indigenous European cultures, bigger Hunter Gatherer types killed off the original farmer males and took over."

How do you know native Europeans were big, angry, stupid, rapist? We don't know this.

a said...

@Romulus
If you could quote your sources for accuracy;for example Cro-Magnon 13k+/- ydna I2a Swiss, Bichon
"60 kg and stood 1.64 m tall" or roughly 5.45+/- feet and slender/gracile build 132lbs+/-
https://en.wikipedia.org/wiki/Cro-Magnon.

Davidski said...

Ryan,

Genome-wide steppe admixture has practically disappeared in many Indo-European groups far away from the steppe simply because their most populous social strata never had that much Indo-European admixture from the steppe to begin with.

Note that today when we compare a skeleton from a high-status grave from what is presumably the early Indo-European period in, say, Scandinavia or Germany, it shows more steppe admixture than the present-day locals.

This can be seen in my latest ADMIXTURE runs, where EHG levels are generally lower in present-day Northern Europeans than among their Bronze Age predecessors, which is something that rk has noticed and commented on above.

However, this often isn't the rule for Y-chromosome haplogroups derived from the Early Bronze Age steppe. They have penetrated much of Eurasia in a big way, even places where we would not be able to detect genome-wide steppe admixture.

Ryan said...

"This can be seen in my latest ADMIXTURE runs, where EHG levels are generally lower in present-day Northern Europeans than among their Bronze Age predecessors, which is something that rk has noticed and commented on above."

That doesn't explain why CHG is so high in those groups though does it.

"Genome-wide steppe admixture has practically disappeared in many Indo-European groups far away from the steppe simply because their most populous social strata never had that much Indo-European admixture from the steppe to begin with."

That admixture run has most Greeks and Armenians on par with Han Chinese, Australian aborigines, Biaka and Mbuti for EHG. That's not practically disappeared. That's entirely disappeared. I get that that ancestry could have been watered down a lot, but I just find it a little hard to believe that there's more WHG in Malays than there is EHG in Armenians. Armenia and Greece aren't exactly the most remote places from the Steppe either.

If it's just a statistical artifact then whatever. I just find it weird is all.

Rob said...

Ryan
I thought sikilar to you initially about the EHG /: CHG thing
By CHG is too broad to be linked with a single language group
Plus we should be careful with modern DNA

I'd say let's wait for more aDNA from Southern Europe, Anatolia and India to make sense of it; so we're comparing oranges with oranges

Davidski said...

Ryan,

CHG isn't synonymous with steppe ancestry in Europe. I think it's pretty clear that there were other sources of CHG admixture in much of Europe.

And Armenia isn't easily accessible from the steppe. It's blocked off by the Caucasus.

But the Bronze Age Armenians do show a lot more steppe admixture than present-day Armenians. If you disagree with what I said above, then what is your explanation for this difference between them?

ryukendo kendow said...

@ Davidski

Thank you very much for the stats! Will plot them later when I have time.

In the meantime, just wanted to chip in that the conclusion that present-day Europeans have much lower levels of steppe ancestry than the LNBA individuals is not yet warranted I believe. They are likely to have a few percents less, but not anything more than >10% less I think, especially for North and Western Europeans.

Generally speaking from periods where we do have genaeological records, elite dominance can produce large levels of replacement even when the elites are very few originally, mainly because of the high heritability of status across generations coupled with nutrition, fertility, and mortality effects on different social classes repeated generation after generation, which we may even have circumstantial evidence for in the form of IE skeletons.

And history has simply not been very kind to hunter gatherers in the post-neolithic period across the world, I really cannot believe that HGs can account for such a high proportion of modern European ancestry independently of HG ancestry in agricultural and pastoral groups, when in other settings they only account for only ~27% of the genomes of Orang Asli for example, i.e. they have been massively replaced even in people with a HG lifestyle themselves.

Not to mention that the new ADMIXTURE still has IE groups 'shifted out' from other Europeans into a cluster of their own, EEFs in another tight cluster, and moderns in yet another very tight cluster, which is diffficult to reconcile with genotype-based PCAs, which show LNBA Europeans more or less squarely within the range of modern European variation. In fact these PCAs show LNBAs and Moderns overlapping greatly in all dimensions that I can remember, from 1-5 or so. So David, I'm still a bit skeptical.

Matt, what do you think?

To get rid of the 'non-supervised bleeding' effects, David, is it possible to repeat the experiment with 2 of each ancient component samples i.e. 2 CHGs, 2 WHGs, 2 EHGs, 2 EEFs, and then running one European plus one LNBA genome at a time, the way you did your original Yamnaya 'flushing'? This should get rid of effects caused by intra-LNBA and intra-European sharing.

Onur said...

@Ryan

That admixture run has most Greeks and Armenians on par with Han Chinese, Australian aborigines, Biaka and Mbuti for EHG. That's not practically disappeared. That's entirely disappeared. I get that that ancestry could have been watered down a lot, but I just find it a little hard to believe that there's more WHG in Malays than there is EHG in Armenians. Armenia and Greece aren't exactly the most remote places from the Steppe either.

What are you talking about? Most Greeks have above noise levels of EHG. Some northern Greeks have as high as 8% EHG.

Davidski said...

rk,

They are likely to have a few percents less, but not anything more than >10% less I think, especially for North and Western Europeans.

I don't understand? Why especially for Western Europeans? Steppe ancestry isn't really all that high in Western Europe.

Ryan said...

@Onur - I'm just looking at the spreadsheet from the K10 run?

@Davidski - I think there's just something going wonky with that run.

I agree re: excess CHG though. It seems to have come after the IE expansion too. Any ideas how/what/why that was? Seems very strange.

Ryan said...

I mean, either there's something wonky there, or PIE is a CHG language, but I think the former is more likely.

Davidski said...

Ryan,

The early to middle Neolithic farmers didn't have much J. The steppe people moving into Europe haven't shown any thus far.

So wouldn't the excess CHG above what came from the steppe be associated with the high levels of J across many parts of Europe today?

Otherwise, if the extra CHG is fake, then where did the J come from and with whom?

ryukendo kendow said...

@ Davidski

I mean that the Steppe ancestry is not likely to have endergone the drastic decrease in N Europeans, falling about a third comparing LNBAs and the moderns from the same geographical locations, that is implied in the current run.

In SE and S Europe the differences are likely to be real, but I think repeating the mass comparison between Nordic LBA and Norwegian, with Sweden BattleAxe and Swedish, late central European LNBAs with modern central Europeans, etc. etc. is not likely to throw up large differences for groups in NE and NW Europe.

Maybe we can try to run repeated sets of datasets comprising two ancient references per component plus one LNBA and one modern? Its no different from the time when you ran Europeans and SC Asians one by one to 'flush out' the Yamnaya in them by preventing Euros and SC Asians from gaining their own components. In this case it may prevent any of the 'anccient' components from drifting around and latching onto stray variation shared by LNBAs and Europeans with each other.

postneo said...

"There's no rule that says ancient West Asia has to be the most advanced. Since the Middle Ages that has not been the case. Everything doesn't have to come from West Asia. No one knew the other existed in ancient times. People lived ignorant of history and geography"

thats always been the rule and has worked like clockwork.

technology and dominance has always been a function of population density and critical mass. Africans hominids and then humans always prevailed over peripheral middle easterns for over 100 k years. then middle easterners acquired critical mass and prevailed over more peripheral humans.

the middle ages is fairly recent in human history when europe acquired critical mass. before that all innovation from agriculture, domestication,immunity , religion, commerce, law, writing, counting, time keeping in europe were adopted from the middle east.

Its also true that the newer initiates prevailed over the more peripheral populations completely but only held temporary sway over older populations. This is borne out by the history of colonialism in the new(americas, australia) vs old worlds(asia, africa).

Onur said...

@Ryan

I'm just looking at the spreadsheet from the K10 run?

There is not any significant difference in EHG levels between K8 and K10, at least for Greeks.

Davidski said...

rk,

Chad is running some stuff with the CHG K10 dataset. Let's wait and see what he comes up with. He's trying to match the results with formal stats.

Postneo,

It makes no difference. The evidence for a massive male dominated migration from the Eastern European steppe to South Asia is now very convincing, and it will be overwhelming next year.

Do yourself a favor and get with the fucking reality already.

Balaji said...

Davidski,

Thank you for the data. Most Indians follow the same pattern as Paniya and Austroasiatic whereby Ust-Ishim does not significantly favor Loschbour. This includes Brahmin_UP who according to Moorjani et al. have 62% ANI as compared to 17% ANI for Paniya. Clearly Indians do not have much LBK_EN or CHG ancestry. The Pakistani populations – Makrani and Brahui - do show evidence of Middle Eastern admixture. They are in Western Pakistan and are known to have Arab as well as African admizture.

Together with the evidence from your Smarter Bear plot, we have conclusive evidence that ANI in India was not brought by farmers from the Middle East but has been in the Subcontinent since the Pleistocene.

The Reich Lab people thought that Middle-Eastern migration was likely. They wrote in Moorjani et al. “The second is that Western Asian peoples migrated to India along with the spread of agriculture; such mass movements are plausible because they are known to have occurred in Europe as has been directly documented by ancient DNA Any such agriculture related migrations would probably have begun at least 8,000–9,000 years BP (based on the dates for Mehrgarh) and may have continued into the period of the Indus civilization that began around 4,600 years BP and depended upon West Asian crops.”

They are now known to be wrong. Instead what Kivisild et al, wrote in 2003 based on mtDNA and Y chromosome data is proving to be more accurate. “Taken together, these results show that Indian tribal and caste populations derive largely from the same genetic heritage of Pleistocene southern and western Asians and have received limited gene flow from external regions since the Holocene”.

Could we do the following D statistics just to get a more complete picture?

Chimp Usi-Ishim Pulliyar Kostenki14_UP
Chimp Usi-Ishim Chamar Kostenki14_UP
Chimp Usi-Ishim Kanjar Kostenki14_UP
Chimp Usi-Ishim Kol Kostenki14_UP
Chimp Usi-Ishim Indian_Singapore Kostenki14_UP
Chimp Usi-Ishim South_Indian Kostenki14_UP
Chimp Usi-Ishim Kusunda Kostenki14_UP
Chimp Usi-Ishim Bengali_Bangladesh Kostenki14_UP
Chimp Usi-Ishim Burusho Kostenki14_UP
Chimp Usi-Ishim Brahmin_UP Kostenki14_UP
Chimp Usi-Ishim Brahmin_TN Kostenki14_UP
Chimp Ust-Ishim Brahui Kostenki14_UP
Chimp Usi-Ishim GujaratiA Kostenki14_UP
Chimp Usi-Ishim Kalash Kostenki14_UP
Chimp Usi-Ishim Makrani Kostenki14_UP
Chimp Usi-Ishim Sindhi Kostenki14_UP

Chimp Usi-Ishim Yamnaya Loschbour
Chimp Usi-Ishim Andronovo Loschbour
Chimp Usi-Ishim Pathan Loschbour
Chimp Usi-Ishim Orcadian Loschbour
Chimp Usi-Ishim Tuscan Loschbour
Chimp Usi-Ishim Lezgin Loschbour
Chimp Usi-Ishim Georgian Loschbour
Chimp Usi-Ishim Armenian Loschbour
Chimp Usi-Ishim Syrian Loschbour
Chimp Usi-Ishim BedouinB Loschbour
Chimp Usi-Ishim Iranian Loschbour

Matt said...

@ Ryu, I guess my opinion is, an excess of extra WHG HG ancestry beyond what would make sense for the 25% in MN farmers mixing with the 0% (or in this even 10% ish) has been found in Haak's qpAdm+resnorm modelling and most of David's ADMIXTURE. There's also a shift in D(WHG,EHG)(Pop,Outgroup) between LNBA Europeans and modern ones. It's a bit of mystery how that happened (how does this happen?), but I think it's at least a consistent finding; it's not novel to this calculator. The changes suggested by this calculator don't seem impossible to me, and it seems improved over K8, that at least the Anatoian Neolithic is the largest single component for most / all Western and Southern Europeans, as it seems like it should be.

@ Romulus, the measurements in Mathieson are good for now. I think there is probably unexplained variance, but whether it will systematically even out what has been found or anything is a question. They're particularly useful in terms of telling us whether there was an overall difference between populations, as opposed to it just being the case that the cultures had different burial traditions for high status people.

Btw. Otzi was 1.65m tall in most references (corpse shrinkage is an issue with measuring this perhaps though). Estimates at changes for males given by http://www.hormones.gr/127/article/article.html are UP: 179cm, 67kg, Mesolithic: 166cm, 62kg, Neolithic LBK 165cm, 64kg, contemporary EHG, 173cm tall. (Direct analysis of the two Barcin males in Mathieson's supplement gave them at 166-168cm which is around consistent with EN, and the supplement doesn't tell us anything about any of the other steppe, MN or LNBA individuals.)

I guess that will do for a reference until anything further develops, or unless there is anything much more systematic sitting in a phys anth textbook.

Re: advancement over time between Europe and West Asia, I guess my take on it is;

During the Upper Paleolithic, areas of Europe were probably more "advanced", from what we can tell from tool pages and that cave art. It might have been harder to do a lot, culturally, for much of the period after that, due to glaciation.

From the Neolithic up to the present, it would've also been hard for Europe to be more advanced, in the sense of having more sedentary culture (opposed to better pastoralist techniques), as crops from Asia Minor would work less well in most of Europe, and microclimates where there work well could've been more spread out (making learning / sharing harder). Still, that didn't seem to stop early and middle Neolithic cultures around the Danube and Balkans (Cucuteni-Tripolye) being arguably more advanced than anything going on in West Asia at the time.

Krefter said...

@postneo,

It's not that simple. I agree with some of what you say though. And about religion, Christianity did not expand in the Roman empire because of any of those reasons. It started out with a handful of people in Israel and expanded within the Roman empire.

Going back to the original discussion about PIE. It's from Steppe/Russia and Ukraine. There's no debating anymore. And we've got to stop using the lack of EHG in India as evidence it isn't from the Steppe. We have documented prove Indo Iranian speakers of Central Asia were descended from Sintahta. Armenia, Turkey, Greece are evidence, I expect a similar scenior in India, with low genetic impact by IEs.

Rob said...

Krefter

"Armenia, Turkey, Greece are evidence, "

I'm curious- can you please elaborate what you mean/ envisage ?

Davidski said...

Balaji,

The conclusions of Kivisild and other scientists from a few years ago are outdated and useless. R1a-Z93 arrived in South Asia from the steppe only around 1800 BC. You will have to come to terms with this sooner rather than later, probably next year. Here are your D-stats.

Chimp Ust_Ishim Pulliyar Kostenki14_UP 0.006 0.787 122979
Chimp Ust_Ishim Chamar Kostenki14_UP 0.0051 0.715 122979
Chimp Ust_Ishim Kanjar Kostenki14_UP 0.0109 1.51 122979
Chimp Ust_Ishim Kol Kostenki14_UP 0.0082 1.156 122979
Chimp Ust_Ishim Indian_Singapore Kostenki14_UP 0.0048 0.768 405033
Chimp Ust_Ishim Kusunda Kostenki14_UP -0.0049 -0.772 549474
Chimp Ust_Ishim Bengali_Bangladesh Kostenki14_UP -0.0036 -0.566 549474
Chimp Ust_Ishim Burusho Kostenki14_UP 0.0008 0.13 549474
Chimp Ust_Ishim Brahmin_UP Kostenki14_UP 0.0109 1.565 122979
Chimp Ust_Ishim Brahmin_TN Kostenki14_UP 0.0114 1.504 122943
Chimp Ust_Ishim Brahui Kostenki14_UP 0.0115 1.882 549474
Chimp Ust_Ishim GujaratiA Kostenki14_UP 0.0024 0.383 549474
Chimp Ust_Ishim Kalash Kostenki14_UP 0.0052 0.825 549474
Chimp Ust_Ishim Makrani Kostenki14_UP 0.017 2.822 549474
Chimp Ust_Ishim Sindhi Kostenki14_UP 0.0038 0.623 549474

Chimp Ust_Ishim Yamnaya_Samara Loschbour 0.0134 2.42 498248
Chimp Ust_Ishim Andronovo Loschbour 0.0092 1.563 499822
Chimp Ust_Ishim Pathan Loschbour 0.0138 2.612 501362
Chimp Ust_Ishim Orcadian Loschbour 0.0163 3.035 501362
Chimp Ust_Ishim Italian_Tuscan Loschbour 0.0205 3.84 501362
Chimp Ust_Ishim Lezgin Loschbour 0.0201 3.696 501362
Chimp Ust_Ishim Georgian Loschbour 0.0207 3.933 501362
Chimp Ust_Ishim Armenian Loschbour 0.02 3.604 501362
Chimp Ust_Ishim Syrian Loschbour 0.0426 7.866 501362
Chimp Ust_Ishim BedouinB Loschbour 0.043 7.987 501362
Chimp Ust_Ishim Iranian Loschbour 0.0226 4.195 501362

Nirjhar007 said...

R1a-Z93 arrived in South Asia from the steppe only around 1800 BC
I hope whoever giving you the unpublished aDNA results are trustworthy mate.

Nirjhar007 said...

We have documented prove Indo Iranian speakers of Central Asia were descended from Sintahta.
Can you reveal them please?.

Rob said...

Dave

Can you post the BA-and IA-Armenian from RISE K8 and K10s please ? I can't see them anywhere on your Xcel

Karl_K said...

@Nirjhar007

"'R1a-Z93 arrived in South Asia from the steppe only around 1800 BC'
I hope whoever giving you the unpublished aDNA results are trustworthy mate."

Exactly what I was thinking. 1800 BC might be a little bit early.

Arch Hades said...

"But the Bronze Age Armenians do show a lot more steppe admixture than present-day Armenians. If you disagree with what I said above, then what is your explanation for this difference between them?"

"Steppe ancestry" as in the Eastern Hunter Gatherer part of the steppe, yes. But I see in the Bronze Age Armenians more European Hunter Gatherer genes than modern Armenians, but less than modern Greeks. So its still not that large.

http://i1067.photobucket.com/albums/u431/ArchHades/Genetic%20maps%20and%20distance%20tables/ADMIXTURE%20Allentoft%20et%20al.%20Supplementary%20data%20cherry%20picked..jpg

Bronze age Armenians sure don't look anything like Andronovo or Sintashta.

ryukendo kendow said...

@ Balaji

Balaji, I disagree. Looking at the stats below, here are all the stats that are significant:

Chimp Ust_Ishim Indian_Singapore Loschbour 0.0134 2.43 368472
Chimp Ust_Ishim Burusho Loschbour 0.0109 2.06 501362
Chimp Ust_Ishim Brahui Loschbour 0.0211 3.903 501362
Chimp Ust_Ishim GujaratiA Loschbour 0.0135 2.317 501362
Chimp Ust_Ishim Kalash Loschbour 0.015 2.661 501362
Chimp Ust_Ishim Makrani Loschbour 0.0272 5.041 501362
Chimp Ust_Ishim Sindhi Loschbour 0.0136 2.563 501362

This list includes even Singapore_Indian, which are middle-caste tamils from the Southern tip of India.

For the stats in the insignificant list, all of them still differ in the correct direction; *but* every single population in the insignificant list is run on less than a fifth of the markers of those in the significant list, and this almost certainly explains why the Kusunda, which are mostly East Asian, as well as Bengalis, are more 'basal' somehow than such populations as the Kanjar, who are Caucasoids from next to Pakistan. In fact, all the populations with Z>1 are run at 500k or 360k markers, while all those Z<1 are run at 100k markers. The ones run at 100k markers look like a parallel cline to the ones on 500k markers but with Z score uniformly depressed.

We also have:
Dai Chamar LBK_EN Kotias 0.0089 3.215 91609
Dai Kol LBK_EN Kotias 0.0096 3.547 91609
Which already show that two of the populations in the 'insignificant' list nevertheless contain Kotias-related ancestry, as there cannot be Dai ancestry in LBK_EN that somehow is not found in Kotias. Notice that this has a large Z score despite being run on 90k markers.

And also:
Dai Paniya LBK_EN Kotias 0.0055 1.742 92265
Dai Austroasiatic LBK_EN Kotias 0.0054 1.962 91505
Admittedly, these only approach significance, but once again note that it only uses 90k markers. I am quite certain these will be likewise removed from the insignificant list once the number of markers increase.

If we repeat the stats with the Brahmin_UP, with Dai used in place of chimp to completely 'subtract' away the ASI portion of their ancestry, I'm quite sure Brahmin_UP will strongly favour Kotias over LBK_EN.


@ Matt

Hmm, about the WHG, I must have misremembered. Looking back at the Haak modelling, it does indeed give up to 15% WHG to some Europeans in the prescence of Yamnaya. Agree that the sociology and anthropology of the 'metaethnic frontier' that resulted in this is kinda difficult to imagine(?) haha.

I still suspect that the overall proportion of Yamnaya ancestry assigned by this ADMIXTURE run is too small. The Haak modelling assigned Bell Beakers ~50% Yamnaya, which more or less completely agrees with their overlap with present-day Northern Europeans in PCA. But the closeness to WHG vs EHG is a good point you raised... Let's see what Chad can bring.

FrankN said...

@Dave: Some notes on your new K10:

1. EHG in SSA: Mota comprises 1.7% EHG. EHG is also found in three Masai_Kinyawa (0.6-1.2%) and three Sandawe (0.04-0.6%). As EHG is rare or lacking in Egyptians, Ethiopians, Somali and Yemenite Jews, this can't be a recent phenomenon (which is also precluded by the EHG occurence in Mota).
Traces of R1b-V88 working its way through Africa? Interestingly, no EHG in Yoruba, so whatever R1b-V88 arrived there seems to be captured in the Afro-Eurasian cluster. This could point at an east-to-west expansion of EHG (R1b-V88 related?) ancestry through Africa.
In any case: Mota as outgroup in EHG-related analyses is likely to distort results (Mbuti, OTOH, might work). Could you include further SSA samples (Ari, JuHoan etc.) for a check how far/deep ancient EHG has worked its way through Africa?

2. WHG in SSA: Similar to the above - 0.2% in Mota, traces in 2 Massai (<0.85%), 4 Sandawe (<0,3%), 2 Datog (<0.85%).

3. 'Oceanian' artefacts in aDNA: Don't know whether its a methodological artefact, or reflects real prehistoric movement, but there is quite some Oceanian admix showing up where I wouldn't expect it. From top to bottom:
- AnatNeol I0725 2.7%
- Altai_IA 2.7%
- Afanasievo RISE510 2.3%
- Poltavka IO126 2.3%
- Iberia_EN IO412 2.3%
- Iberia_MN IO407 2.2%
- Satsurbia 2.1%
- CW Germany I1542 2.1%
- BB Germany RISE559 1.9%
- Andronowo RISE503 1.9%
- Poltavka IO374 1.8%
- LBK_EN I0659 1.6%
- Hungary_BA I1502 1.5%
- Samara_Eneol I0122 1.5%
- Okunevo RISE515 1.5%
- Yamn_Samara I0439 1.4%
- Esperstedt_MN 1.4%
- CW Germany RISE446 1.3%
- Hungary_EN I1495 1.3%
- Nordic_LN 1.3%
- LBK_EN I0057 1.3%
- AnatNeol I1098 1.3%
- Mota 1.2%
- Karsdof_LN 1.2%
- Hungary_EN I1499 1.2%
- Kotias 1.1%

I stop here, but there are many more. 'Oceanian' ancestry above 0.01% is shown for:
- 57% of "Steppe" aDNA samples
- 48% of Neolithic aDNA samples
- 48% of Chalcolithic/BA samples
- 95% of current South-Central Asia samples
- 87% of current Caucasian samples
- 81% of current Near Eastern samples
- 70% of SEA (incl. Han/Japan) samples
- 56% of current Med/Balkan, and 42% of Basque/Sardinian (not included in the above) samples
- 48% of Siberian samples
- 38% of non-Med/Balkan/Caucasus European samples
- 24% of Sub-Saharan Africa samples.

Whatever strange element we are dealing with here - it is neither related to EHG and CHG (already present with Early Farmers), nor to Anat_Neol (present with Afanasievo, Okunewo etc.). There is, however, some tendency to vary with the WHG share. Could you include Nordic stuff (Motala etc.) to check whether they also display 'Oceanian' admix?
Otherwise, one more warning against Mota as outgroup. Oceanians should also fall from that list, as should Dai (all shown as 'oceanian' admixed).

Chad Rohlfsen said...

EHG in Mota reflects his ancestry being closer OoA. The LN/EBA have about 70% coverage on average. There's only about 10 that are excellent. Spain EN looks like they have a bit of their own drift. Look at a pebble and you'll miss the mountain.

Chad Rohlfsen said...

I'll see if I have something to share tonight. It's still a bit rough. I may have it completed tonight.

Matt said...

@rk - Hmm, about the WHG, I must have misremembered. Looking back at the Haak modelling, it does indeed give up to 15% WHG to some Europeans in the prescence of Yamnaya.

In a way I prefer to look at the EHG models at the moment, as I'm a bit wary of their Yamnaya models in a way, as although it seems like most CHG should have come from Yamnaya, you really want to allow EHG and CHG to vary freely and see what happens (which it seems to be hard to find adequate outgroups in qpAdm for, from Davidski's comments).

(I'm also not sure the Haak modelling there can cope so well with any Yamnaya plus Nganasan together well, but that's another topic).

So I'd prefer to look at the EN+EHG+WHG models, as a substitute for Anatolia_EN+CHG+WHG+EHG models.

Just to illustrate about the Haak modelling again this, taking a couple examples from the Haak modelling supplements:

English - 0% BedouinB +17.9% EHG + 63.8% EN + 1.4% Nganasan + 16.9% WHG

Assume EN is as Mathieson et al describes it and approx 7-11% WHG (go for 9% as a median).

Then English - 0% Bedouin + 17.9% EHG + 58% Anatolia_EN + 1.4% Nganansan + 22.4% WHG.

Or with the EHG+EN+WHG model (which is about exactly as good a fit for English)

English - 18.6% EHG + 57.5% EN + 23.9% WHG
goes to English - 18.6% EHG + 52% Anatolia_EN + 29% WHG

Davidski's model gives English - 1% SW Asian + 15% EHG + 57.5% "Neolithic" (31% Anatolian_EN + 26.5% CHG ) + 0% ENA + 25% WHG.

Not far from the same neighbourhood, assuming that qpAdm+resnorm (Haak modelling) can't really grok the difference between CHG and Anatolia_EN so easily based on the outgroups Haak et al or we are feeding to it.

Another example that might not work quite so well -

Belarusian - 0% BedouinB + 19.5% EHG + 48.3% EN + 2.3% Nganansan, 29.8% WHG
goes to Belarusian - 0% BedouinB + 19.5% EHG + 44% Anatolia_EN + 2.3% Nganansan, 34.15% WHG

or Belarusian - 20.6% EHG + 38.2% EN + 41.2% WHG
goes to Belarusian - 20.6% EHG + 34.8% Anatolia_EN + 44.6% WHG

vs Davidski's Belarusian - 1.4% SW Asian + 18% EHG + 51.1% "Neolithic" (23.9% Anatolia_EN + 27.2% CHG) + 1.3% ENA + 27.6% WHG.

Davidski's K10 actually maybe underweights the Haak WHG fraction here. Still seems fairly close I guess though?

Grey said...

Kriistina
"However, it cannot be the whole story as wagons are quite useless in a forest, and IE languages spread for example to Scandinavia where wagons were not very practical."

Maybe the very first "wagons" were Travois.

https://qph.is.quoracdn.net/main-qimg-e6253060876bd8913729ec3a19f6b9ab?convert_to_webp=true

And then someone stuck two wheels on the end of the poles?

Early linguistic confusion over "wagon" and "chariot" that eventually sorts itself out might hint at something like that.

.

Romulus

"So the Sredny Stog people traded their seashell pots to the Tripolye people for metal objects."

If the movement of females was mostly A to B then generally that would imply B had the valuable trade items.

.

As a "what if" if Kargaly was the origin how could a Kargaly-origin model fit the data...

Kargaly copper field
-> locals produce simple copper objects like beads
-> long distance trade in these objects
-> uses up a lot of wood
-> steppe copper workers run out of wood (in the quantities needed)
-> steppe copper workers move to regions known to have both copper and wood
-> leads to bronze off the steppe

this model could potentially explain
- early R1b spread
- later R1a turnover from yamnaya-afenevso to cordedware-sintashta

i.e. (in the far north) R1b had the copper first, R1a had the bronze first

FrankN said...

@Chad: "EHG in Mota reflects his ancestry being closer OoA."

In that case it should be more widespread around the Horn of Africa. But EHG is missing, e.g., from Hazda, all Somalis except one, and all Eth_Tigray and Eth_Jew. O.k. - Amharan migration sometimes around 300 BC should have diluted previous EHG, but not completely replaced it.
As i said - would be good to have more Sub-Saharan African Samples included to see if patterns are showing up.

Anyway - Dave's K10 shows EHG with a couple more samples where its presence is difficult to explain by OoA, or some 'kurganist' expansion:
- Kotias 5.2%
- Anat_Neol I0727 5.1%
- Anat_Neol I0725 4.1%
- Hungary_EN I1499 3.9%
- Hungary_EN I1498 3.3%
- Loschbaur 3.2%
- Anat_Neol I0724 3.1%
- LBK_EN I1550 2.9%
- Anat_Neol I0746 2.6%
- Hungary_EN I1506 2.3%
- Anat_Neol I1103 2.0%
- Hungary_EN I1500 1.7%
- Anat_Neol I1580 1.6%
- LBK_EN I0025 1.5%
- Hungary_EN I1495 1.2%
- Hungary_EN I1496 1.2%
- Anat_Neol I1102 1.1%
- Anat_Neol I1099 1.0%

plus, again, a couple more (early) neolithic samples from Anatolia, Hungary and LBK, plus Iberia_MN, Baalberge_MN, and the Iceman.

Looks like EHG was quite western (Loschbaur) and also quite Caucasian (Kotias) already during the Mesolithic, and readily absorbed into the early Danubian Neolithic. At 1.7% average EHG in Hungary_EN, and 1% in Anat_Neol, it shouldn't take much more EN sampling east of the Rhine before we get the first Neolithic R1b there.
And afterwards, the whole IE discussion is "back to square one". Well, maybe "Square 2.5". "Out of Anatolia with early farmers" should be dead, and the "South Baltic to Pontic and Caspian shores"-rhumbus (which in parts is Steppe, in other parts mountains, forest or forest-steppe) definitely played an important role.

ADDENDUM: CHG averages:
- Anat_Neol 2.0%
- LBK_EN 1.3%
- Hungary_EN 0.7%

Highlights:
- Anat_Neol I0708 12.8%
- LBK_EN I0048 7,2%
- Anat_Neol I1098 7.2%
- Anat_Neol I0726 6.3%
- Anat_Neol I0707 5.6%
- LBKT_EN I0176 4.8%

The mix has culminated somewhere between Yamnaya and Eastern CW, but it definitely didn't start there, but was already present in LBK and on the Balkans and/or West Anatolia. High time we get CT aDNA!

Chris Davies said...

@ FrankN - re: EHG in SSA - R1a has turned up in several SSA samples, including Somalia, Mozambique, Zambia, Namibia, Sao Tome. It could be all colonial-era, or..

Chad Rohlfsen said...

Motas stuff is noise. It's not to be taken literally. He is full SSA, that shares more drift with Eurasians than what is found in the San, Mbuti, or Yoruba, whom are all admixed with real West Eurasian ancestry.

Grey said...

Matt

"(So what does "massive Protoeuropoid type" mean here? Large body / large face?)."

specifically seems to mean heads/faces

http://s48.radikal.ru/i121/1108/37/7d43a2b33edct.jpg

but more generally

https://en.wikipedia.org/wiki/Surface-area-to-volume_ratio

https://en.wikipedia.org/wiki/Bergmann%27s_rule

"Bergmann's rule is an ecogeographic principle that states that within a broadly distributed taxonomic clade, populations and species of larger size are found in colder environments, and species of smaller size are found in warmer regions."

something to do with heat loss

"In particular, reversible dwarfing of mammals has been noted during two relatively brief upward excursions in temperature during the Paleogene"

So...maybe mega hominids that went with the mega fauna and then shrunk as it got warmer.

Grey said...

ryu

"I really cannot believe that HGs can account for such a high proportion of modern European ancestry independently of HG ancestry in agricultural and pastoral groups, when in other settings they only account for only ~27% of the genomes of Orang Asli for example, i.e. they have been massively replaced even in people with a HG lifestyle themselves."

Looking at it logically a model that would explain it might be

1) a region of Europe that was still heavily WHG because farmers had a hard time in that region for some reason for example
- acid soil
- wheat being low yield in acid soil

combined with

2) a heavily male mediated migration from elsewhere who
- married whg heavy local women
- brought with them some way of compensating for the acid soil

for example cattle raising and lactose tolerance

such that

3) the combined whg heavy and lactose tolerant population expanded dramatically

Ryan said...

@David - I don't think the extra CHG is fake. I just think it may be inflated (presumably taking some of EHG?)

Re: J - J2 has a coastal distribution in Europe doesn't it? Can we say that about the excess CHG?

And are you including J1 too? If so, what would the "original" Levantine and Arabian haplogroup be?

And what I mean more about who these people are - it looks like the CHG expansion occurred late enough that it should have left a historical record. So I'm asking if there's a candidate for this group.

Re: Oceanians - David got some wierd Oceanian connections for La-Brana, so maybe it's a real signal. A relic from the first OoA migration.

Grey said...

postneo

"technology and dominance has always been a function of population density and critical mass"

I think that is generally true and that after farming was developed and after the tools necessary to drain big river valleys were developed the "technology and dominance" would have shifted to the big river valleys.

However before that it seems to me

- technology would initially develop among people who happened to live close to any needed resource

- farming would initially develop among sedentary HGs who happened to live near an animal or crop that could be domesticated

and if so then the *initial* hot spots are likely to be scattered all over the place.

Grey said...

One last "what if" on large size and temperature - if correct and if EHG was connected to that then it might have been selected against as it got warmer i.e. maybe when there were mammoth to feed on there was lots of Klitchkos but after taking up farming being that big - although useful in some ways - became a net liability.

capra internetensis said...

Thanks for all the goodies, David.

Looking at the CHG/EHG ratios of Bell Beaker and Corded Ware in this K10, they have hardly budged from where they were in K8, except that Corded Ware RISE436 has lost about 10% CHG and gained 10% SW Asian, bringing his CHG/EHG ratio down from 1.8 to 1.2. He has pretty low coverage though. Actually most of the samples have poor coverage, but the ratios don't seem to vary with coverage; none of the really good samples has CHG/EHG greater than 0.9, though.

Some of them have gained greater than noise levels of SW Asian, but without any obvious pattern. There are two Bell Beaker men with ~7% SW Asian, but there doesn't seem to be anything unusual about them otherwise (both are from the same place and time and have mtDNA H1, so probably from same population though not close relatives).

There isn't any consistent connection with CHG though, it's yet another independent component.

Tobus said...

@ryukendo:The ones run at 100k markers look like a parallel cline to the ones on 500k markers but with Z score uniformly depressed.

Fewer markers will depress the Z-score which is a sum, but should have little effect on the "D-score" which is a percentage. A D-score of ~0.01 or greater is usually accompanied by a significant Z-score, so extrapolating from this I'd say these two results are not overly significant:
Chimp Ust_Ishim Pulliyar Kostenki14_UP 0.006 0.787 122979
Chimp Ust_Ishim Chamar Kostenki14_UP 0.0051 0.715 122979

...but these ones are:
Chimp Ust_Ishim Kanjar Kostenki14_UP 0.0109 1.51 122979
Chimp Ust_Ishim Brahmin_UP Kostenki14_UP 0.0109 1.565 122979
Chimp Ust_Ishim Brahmin_TN Kostenki14_UP 0.0114 1.504 122943

The other low-coverage pop is probably borderline:
Chimp Ust_Ishim Kol Kostenki14_UP 0.0082 1.156 122979




Rob said...

@ Ryan

As someoone who also looks for alternate models, or at least probes existing ones, maybe for the now we should rest on Occam's razor:

* CHG spread in the north with R1b (and lesser extent) R1a; from the steppe. However they got it, they did. OF course, it is theoretically possible that CHG entered northern Europe independent of the steppe - and it likely did later- but for the initial part I think it simply came with eastern Euro-derived factions of the BB network.

* In southern Europe it probably arrived in several waves (but the largest one between 4000 - 2500 BC), and i suspect in different sub-forms. I suspect that once we resolve J2 better, we might see different patterns between j2a-M67, J2a-Page55, J2b, etc groups. Some of these might have been southern IEs (eg certain J2b coming with R1b-Z2013).

J2's distribution in modern day Europe is n;t 'coastal' only. It has frequencies of 5% up to Hungary and Ukraine. This might not seem like much, but that's probably becuase there's been significant turnover since the Bronze Ages, and much of this has been in the way of north to south (Celts, Slavs, Germanics), even if the overall trend has been a 'Danubianization' of north European groups.

Alberto said...

Going to those Haak et al. models using EN, MN, Yamnaya, EHG, WHG, Motala,..., I also remember that Unetice was best modeled as some 55% Motala and 45% MN. While I don't think that's totally correct, D-stats did confirm that Unetice had higher affinity to Motala than to WHG or EHG. So some good amount must be SHG-like. The stats showed a similar pattern with Lithuanian, IIRC.

I suspect that those Hungarian BA samples with high European HG ancestry might also be best modeled using Motala.

Scots and Tuscans got the worse scores after adding all populations. I always thought they had extra "teal" over Yamnaya, so maybe adding CHG would have worked best for them.

Rob said...

Alberto

I suspect a belt of "Motala"-like groups existed in that vast unsampled region of east -Central Europe: from the Baltic to the east Carpathians

Shaikorth said...

Tobus, looks like reducing the number of SNP's can have a significant effect on D-score too.

Anatolia_Neolithic Armenian Mota Primate_Gorilla -0.0085 -4.2 480195

Anatolia_Neolithic Armenian Mota Primate_Gorilla -0.0047 -1.858 114965

FrankN said...

K10 comments #3

1. 2% Sub-Saharan admix in Sintashta, Okunewo, Nordic_BA? 1.3% SSA in Kotias? 10% of "Steppe", 13% of Neolithic and 19% of Chalcolithic/BA aDNA with more than 0.5% SSA admix? All Australians and Austroasiatic_Ho, and 35% of Tuvinians clearly SSA-admixed?
I am the last one to deny the possibility of long-range migration already during prehistory, but this looks like an artefact. Something that should be Eurasian admixture in Yoruba and Australians (possibly of South Asian origin) is instead showing up as SSA admixture in Kotias etc. I think the SSA baseline needs to be re-calibrated, without Yoruba, and using other populations (but which ones, aside from Biaka/Mbuti?) instead.

2. SE Asian: Kind of the same. Here a list of those that look strange to me:
- CW_Germany I1538 2.6%
- Hungary_EN I1506 2.1%
- BB_Czech RISE568 2.1%
- CW_Germany RISE446 1.94%
- Satsurbia 1.8%
- CW_Germany I0106 1.6%
- Iberia_MN I0405 1.6%
- BattleAxe RISE94 1.3%
- Iberia_Chalc I1282 1.2%
- BB_Germany I0111 1.2%
- Kotias 1.1%
- Alberstedt_LN 0.9%
- Anat_Neol I0723 0.8%

plus a couple more BB, CW, and MN/Chalcolithic Iberians with SEA Admix >0.3%.
In Eurasia outside SEA and NE-Asia, the SEA-Admix tends to go alongside CHG, at a rate between 2-5% (Caucasus/ Near East) and 30% (Altai) of CHG. Within Europe, it seems to peak along a line from Belorus to Iberia, with decreasing shares towards Scandinavia/Iceland and Italy, and "SEA-Admix" absent from Cyprus and Sardinia. Looks like SEA captures some minor contribution to the "teal" that is in fact more related to CHG; though somehow different from it, possibly originating rather from (North-?)east than west of the Caspian sea.

@Chad: "re: EHG in SSA - It could be all colonial-era"
If we were talking Somalis here - yes, it could be. But the EHG admix is within Sandawe and Massai. The top ranks in German East Africa were of course filled by Prussian nobility, who might have 'contributed' some R1a. But otherwise it was mostly marines and merchants, Hamburg/Bremen/Lower Saxony/Schleswig-Holstein, old Anglian-Saxon* stock, predominantly yDNA I1/I2/ some R1b, similar to what could be expected from British 'efforts' in that matter. Anyway, the EHG admix is autosomal DNA, which requires a bit more than colonial-era relations to change.

or..Motas stuff is noise.Might be. In that case we are all in trouble, because their only seems to be a handful of aDNA that isn't 'noise'.
Alternatively, Karelia_HG and Samara_HG may not be the best/ most complete representation of EHG. They seem in fact be drawn towards WHG (note the 3.2% EHG in Loschbaur and other phenomena described above). So far, we don't have anything better. But if we had, quite some of that 'noise', not only in Mota, but also the equally noisy patterns of SEA, SSA and Oceanian admix, might eventually disappear.

*) The Saxons are those with last name on -s, Anglians & Jutes have their last name on -sen. So, "Rohlfs" would be Saxon, "Rohlfsen" Anglian/ Jutish.

Davidski said...

SSA admix in the ancient samples is post-morten deamination (DNA damage), and it shows up more clearly in the RISE samples because they weren't UDG treated.

Davidski said...

New CHG K10...

https://docs.google.com/spreadsheets/d/1Gn0Bea98xDJewbBmh1Yua7J6KH9uPAwrGNQqpwmHvtA/edit?usp=sharing

Rob said...

Where are BA Armenians ?

Davidski said...

Here they are...

Pops Ids SE_Asian Anatolia_Neolithic Caucasus_HG Eastern_HG Afro-Eurasian NE_Asian Sub-Saharan Oceanian SW_Asian Western_HG
Armenia_BA RISE396 0.02541 0.112346 0.54356 0.110403 0.00001 0.00001 0.001127 0.02384 0.132542 0.050753
Armenia_BA RISE397 0.038389 0.140552 0.545596 0.093222 0.00001 0.00001 0.023608 0.001124 0.1086 0.048888
Armenia_BA RISE407 0.035231 0.167554 0.519503 0.134068 0.000841 0.00001 0.013263 0.002809 0.069086 0.057635
Armenia_BA RISE423 0.027141 0.155362 0.57563 0.084436 0.00001 0.00001 0.021979 0.018316 0.074057 0.04306
Armenia_IA RISE408 0.00001 0.261005 0.562623 0.086651 0.00001 0.019722 0.00001 0.02266 0.00001 0.047299

Rob said...

Thanks Dave
So the new "SW Asia" component just slightly reduces EF_Anatolia score ?

Ariele Iacopo Maggi said...

A quick reflection, there is an orcadian with 2% SW asian and a Sardinian with 5% SW asian, to me that fit the proportion of their EEF ancestry. Are you all still sure that CHG came in Sardinia recently? Palestinian have some 30%, cypriots some 20%. On my opion nothing came to island from the middle east in the last 3,4K years.

Balaji said...

Davidski, thanks. Tobus, thanks for the information on D-statistics.

Ryukendo Kendow,

I don't see exactly where you disagree with me. I had already noted the following statistics.

Dai Paniya Kotias Chimp -0.0132 -3.247 113679
Dai Paniya LBK_EN Chimp -0.0092 -2.997 134550

These indicate that even Paniya with only 17% ANI, nevertheless is preferred by by Kotias and LBK_EN over Dai. I am certainly not saying that there is no Kotias related or LBK_EN related ancestry in India. We know from ADMIXTURE that there is a Gedrosia component in India that is related to the Caucasus component in the Caucasus, Jones et al. also showed that populations in India have a large outgroup f3 shared drift with Kotias (Supplementary Figure 3). They also found that the most negative f3 statistics for Indian populations involved Kotias (Supplementary Table 9).

My point is that the Kotias-like component in India is still different from actual Kotias, It is also different from LBK_EN. This was already suggested by ADMIXTURE. It is also suggested by the Smarter Bear plot.

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

Additional support comes from D statistics.

Chimp Ust_Ishim Brahmin_UP Loschbour 0.0067 1.101 112450
Chimp Ust_Ishim Georgian Loschbour 0.0207 3.933 501362
Chimp Ust_Ishim Orcadian Loschbour 0.0163 3.035 501362
Loschbour Kotias Ust-Ishim Chimp 0.0305 4.665 404058
Chimp Ust-Ishim LBK_EN Loschbour 0.0151 2.692 279178

Since the “basal” component in India is so different from that in the Near East and in Europe, it could not have been brought by agricultural immigrants but instead must be indigenous.

V.R. said...

Balaji,

From where are you getting that Paniya are only 17% ANI? Is this certain?

Chad Rohlfsen said...

Those numbers are incorrect. Paniyas are about 40-45% West Eurasian.

Balaji said...

The ANI estimate for Paniya is from Moorjani et al. You can read the paper here.

http://www.cell.com/ajhg/abstract/S0002-9297(13)00324-8

ryukendo kendow said...

@ Balaji

But balaji, I have already highlighted that kusunda and bengali cannot possibly be more basal eurasian than kanjar from rajasthan and gujarat. Neither can orcadian be more basal eurasian than LBK EN, which is what a naive reading of the stats you posted will have us conclude. The fact that brahmin UP scores low in that comparison is almost certainly an artifact from the low number of markers. All the other stats you posted have multiple times the markers of the brahmin UP comparison, and are therefore more significant.

The Singapore Indian comparison gives us a z score of 2.4 at 360k markers. This is a very nice signal of 'normal' basal eurasian ancestry in a population in the far south of India.

I reserve judgement on whether there is yet another branch of basal eurasian in the south of India, there very well might be, but the stats you posted do not support this; neither can we say that this ancestry must be here in the pleistocene from just the numbers. Far more favorable to your hypothesis is in fact some other stats in a previous batch, which showed how difficult it was to get s Indians to favor kotias over lbk even after dai ancestry is subtracted.

James said...

It's so obvious that PIE comes from CHG and not EHG. Only someone desperately wishing for something advanced to originate in Eastern Europe could fail to see that. It's the "teal" component. It's Dienekes' "West_Asian" component. It's why western and southern Europeans have more CHG than EHG.

http://dienekes.blogspot.com/2015/11/westasian-in-flesh-hunter-gatherers.html

It's also why CHG stretches all the way to South Asia with no EHG. India's ANI component shares the most drift with Georgians and other Caucasus groups, and Brahmins are high in J2.

http://dienekes.blogspot.com/2013/08/major-admixture-in-india-took-place-42.html
http://dienekes.blogspot.com/2005/11/more-on-r1a1-age-and-haplogroup-j2-in.html

The steppe was just one place that received IE languages thru CHG and then spread them further west thru Yamnaya, but many other places received them directly from West Asia with metal working.

http://dienekes.blogspot.com/2012/07/a-physico-anthropological-study-of.html

This also would explain the proposed relationship between IE and AA languages.

http://en.wikipedia.org/wiki/Indo-Semitic_languages

Onur said...

Which of the analyses are prone to effects of drift? Formal analyses, ADMIXTURE or PCA?

Davidski said...

James,

Nonsense.

Most of the CHG and J in South Asia arrived there with Neolithic farmers, possibly Dravidian speakers.

That kills your CHG = PIE theory, because Indo-Europeans move into West and South Asia from the steppe later.

This matches historical linguistics evidence. Your theory is as loopy as Renfrew's dead Anatolian hypothesis.

Kristiina said...

Thanks for new CHG_K10!

In CHG_K10 Estonians have the following averages: 20-23% Anatolian Neolithic, 22-26% Caucasus HG, 18-22% EHG, 30-31% WHG, 0-5% NE Asian, 0,7% Oceanian, 0,8% SW Asian
Corded Ware Estonia: 24% Anatolian Neolithic, 22% Caucasus HG, 32% EHG, 21% WHG, 0-0,8% Oceanian, 0-1,5% SW Asian

It means that WHG has gone up significantly while CHG is not any lower and EHG has gone down significantly. I would conclude that in modern Estonians there is a significant WHG continuity from the Mesolithic. My impression is that there is a WHG reservoir around the Baltic Sea. In K8 (before discovery of CHG), my WHG percentage is 50%. Therefore, there must be a significant hunter gatherer gene survival around the Baltic Sea but this should be nothing new. It was confirmed already in Haak et al in February.

@Der Frost “Differences between anthropological types of males (robust UP) and females (gracile Mediterranean) belonging to Late Chalcolithic / Early Bronze Age cultures of Romania-Moldova-Ukraine”
That difference can only last if there are new males coming from robust type areas and new females coming in from gracile type areas, as it can very well happen that a robust masculine man who falls in love with a feminine gracile woman have a gracile boy and a robust daughter :-), although in the long run the averages should prevail.

Shaikorth said...

"Der Frost “Differences between anthropological types of males (robust UP) and females (gracile Mediterranean) belonging to Late Chalcolithic / Early Bronze Age cultures of Romania-Moldova-Ukraine”
That difference can only last if there are new males coming from robust type areas and new females coming in from gracile type areas, as it can very well happen that a robust masculine man who falls in love with a feminine gracile woman have a gracile boy and a robust daughter :-), although in the long run the averages should prevail."

That is, assuming the men and women from those cultures are autosomally different from each other in the first place. I wonder how that will turn out.

Matt said...

Re: New K10 -

It seems like the FST distances make a lot more sense in this one than K8:
PCA of FST distances in K8 vs K10 http://i.imgur.com/Qlo9nJ6.png

It seems like too high FST / otherwise strange distances of WHG in K8 gave a different shape, while K10 seems more to have an expected shape.

(Also in table form attempt to adjust the FST distances to the same level of drift, using SS as the outgroup, to see which should be truly closer to one another, phylogenetically, when drift is taken into account http://i.imgur.com/tt25RVz.png )

The "to component" FST based PCA seem as good as ever, although the SW Asian component seems to have a more dominant character that squeezes some of the Sardinian-Basque-EEF populations together a bit more http://i.imgur.com/oEsbDKs.png

(same thing for the adjusted FST distances table above - http://i.imgur.com/0M49ub7.png).

Onur: Which of the analyses are prone to effects of drift? Formal analyses, ADMIXTURE or PCA? Formal stats in general should be less affected by drift differences between phylogenetically identical populations. At the same time, ADMIXTURE and PCA should be less affected by drift which is particular to particular populations the more populations the more different populations are included (unless it's very high and old). So useful unless that's what you actually want to know about.

FrankN said...

@Dave: "New CHG K10..."

Thanks, Dave. I am in the process of looking at what has changed.

As Christmas is approaching, here already my "wish-list" to be considered in eventual further updates:
1. Could you add a third descriptive column with sample classifiers? They greatly facilitate screening for outliers and general trends via pivot analysis/ selected sorting. For orientation, these are the classifiers I am using so far - tentative, open for review, and ultimately of course your decision:
- aDNA: "HG", "Neol", "Steppe", "Chalc", "IA". "Chalc" includes LN (CW, Nordic_LN etc.) and BA - essentially anything affected by the major CHG/EHG entry into Northern-Central Europe. Haven't yet decided how to assign the Armenian aDNA - possible a class of its own.
- Outside Europe: "Amerind", "SSA", "Ocean", "SEA", "Siberia", "SCA", "NearEast". Mostly self-explaining.
SSA includes Mota. Might be further divided into "EastAfr" and "other SSA", but so far the larger grouping has shown to work for its purpose, i.e. identifying unexpected admix.
SEA includes Han and Japanese, which may alternatively be shifted into the "Siberian" group.
SCA is essentially everything speaking Indo-Aryan, starting with Kurds (but Ossetians are in "Cauc").
Delimiting Cauc and NearEast is somewhat fluent. So far, all of Turkey except Trabzon is in Near East, but that may be reviewed.
- Europe: "BasqSar", "Med", "CEU". "BasqSar" is in future meant to identify possible specific patterns. Haven't really started that, so don't know whether the separation makes sense.
"Med" is all groups/ countries bordering the Mediterranean, including France_South, all of Iberia except Basques, and Bulgaria / Romania (the latter two up for future review). Cyprus might be better placed with NearEast.
CEU is everything else west of the Urals, including Mordwinians etc.
- Jews have been assigned in line with their country of origin.

2. Additional samples:
- Hungary_IA would be great to trace possible secondary CHG incursion via the Balkans during MBA/LBA. A look at PCAs that include it suggests quite a 'caucasian' shift compared to Hungary_BA.
- Some more SSA stuff, most notably JuHoan, but also, if available, West/ Central Africans that are less suspect of prehistoric Eurasian admix than Yoruba.
- A bit more North-Central European DNA (Swedes, Danes, Germans, Dutch, Belgians, Swiss, Austrians), even if its only 3-4 samples per country. So far, some 50% or more of aDNA isn't matched by any current DNA from the same region, which doesn't really help with identifying population trends.

Shaikorth said...

Matt, have you read this:

http://biorxiv.org/content/early/2015/12/08/033852

It's a very good paper when it comes to demonstrating the shortcomings of fst, although their est-method is not yet easily available.

https://pbs.twimg.com/media/CVz7X4kUwAAFMbL.jpg:large

Est correlated better with formal stats than fst:

http://oi68.tinypic.com/2zeausi.jpg

Also correlates better with IBS, with same SNP's if we compare HGDP (Kargopol) Russians and French to something like Yakuts, the French have about 10% higher fst distances (ADMIXTURE tends to reflect this too) but just close to 1% less IBS sharing.

epoch2013 said...

@Kristiina

"Agricultural terminology is not shared to a great extent as I could find only the word for quern."

Mallory says otherwise and considers that a problem. Read about it here, and check table 2:

http://jolr.ru/files/%28112%29jlr2013-9%28145-154%29.pdf

Chad Rohlfsen said...

Here's a supervised K11. I selected a N African component to show the flow into Europe. I'm doing some unsupervised runs right now.

https://docs.google.com/spreadsheets/d/1cihUQZAykzy2kFEr9-mv6RRPxry2PHWDjkgpVDdOJ_Q/edit#gid=729830007

Matt said...

@ Shaikorth thanks. Very interesting.

To try and dumb it down, the EST measure seems to be a here a measure of population structure.

FST is a measure of population differentiation, that is strongly effected by genetic diversity within and between population.

That doesn't necessarily provide a clear indication of the population structure, which the paper more or less describes (helpfully) as a departure between panmixia between the theoretical populations together with panmixia within each of the theoretical populations.

The EST measure is based on differences in SD (dispersal) between the populations and the merged overall population including both pairs rather than differences in means.

I would not agree the EST is a superior measure overall necessarily though as you might want measures of how much actual drift and isolation and differentiation is involved in a population separation, rather than the level of structure only, since you can have wildly different levels of this for similar levels of structure. Patterns of evolutionary selection can follow the level of genetic diversity (which makes sense to me as the diversity to a degree controls what there is to select on), and it's not clear to me that looking only at structure will necessarily provide better signals of long term selection.

However EST looks maybe superior for actually quantifying population structure, as opposed to a sort of "With these populations considered, ADMIXTURE says these are separate clusters" which isn't very good and has many conditions.

In terms of the specific population comparisons, looking at the EBT measure in the supplement -

EBT is said to be is similar in many ways to EST, though its HGDP ranking order is often intermediate between FST and EST" and "an additional equidistance index, denoted EBT, which is less sensitive to intra-population structure and the inclusion of relatives".

EST is noted in the paper as being sensitive to structure within the populations involved, and this will decrease EST ("While FST is weighed down by high diversity within populations, EST is weighed down by high structure within populations.").

Possibly interesting that there is strong EBT separation between the Naxi with Yi and Lahu groups, and weak EBT separation between Cambodia-Mongolia (and also Japanese and Chinese South)?

(As they note though, SD is hard to measure so requires larger SNP counts.)
It could be interesting to systematically compare the measures between populations, to identify population splits that involve high FST (diversity and frequency differentiation) vs EST / EBT (structural differentiation) -

e.g. using the EBT measures
http://i.imgur.com/pPIwFBF.png- All Pairs
http://i.imgur.com/WwXehG8.png- West Eurasian Pairs
http://i.imgur.com/cR2jHJE.png - East Eurasian Pairs

I noticed they also had a EST median measure as well as the EBT... so same pairs with those All - http://i.imgur.com/Rt3QAHu.png, WE - http://i.imgur.com/tibtUvo.png, EE - http://i.imgur.com/1aPtysJ.png. This changes the rank order of structuring (but still for example shows higher structuring between Naxi-Cambodian and Naxi-Yi than Naxi-Mongola). I didn't bother with the EST mean measure.

Shaikorth said...

Naxi distances are distorted becausue the sample has relatively low base-level population structure but has two close relatives in it, resulting in pairs of individuals having higher SD. See figures s3-s5.

Alberto said...

@Chad

That admixture run looks very solid, thanks.

It's interesting that here the CHG component which is maximized in both Kotias and Satsurblia at 99.99% peaks in the Kalash at some 61%, followed by Brahui, Balochi, Sindhi, Makrani, Gujarati, Pathan... Abkhasian and Georgian also appear at these levels 52-55%. It would be interesting to see the results of IBS from Kotias and/or Satsurblia to check how much they differ from shared drift with formal stats, and how much they agree with this Admixture. I do think that it's correct that the Kalash might have the highest CHG, though I don't know if it will show up with Kotias or we'll need samples from further east.

Also interesting the results for Sweden_NHG. Finally we see clearly a solid 20% Anatolia_Neolithic admixture in them. Is it a similar case as with EHGs in the steppe? Getting admixture from nearby farmers? Also female mediated? (I don't think these had any G2a or other "farmer" Y haplogroup). Also should be noted that even with the 20% Anatolia_Neolithic they still score 30% EHG (Motala would be like 40%). This is important to calibrate the "steppe admixture" too (now that we've seen the dramatic effect of CHG being tested independently from EHG or Yamnaya). Some pre-CW DNA from around Poland would be nice for this.

For S-C Asia a modern (Mesolithic) ANE genome from around the Altai would be good too. Most of the EHG there is actually ANE, so that would calibrate things too. Okunevo is some 50-60% MA1, so it's possible that there were pure ANE guys around a bit earlier.

When ordering by EHG, it's quite amazing how much the ancient samples appear at the top. I guess because most of the samples we have are the steppe and steppe derived ones (we really lack DNA from roughly all the relevant places at that time). When it comes to modern populations, it's as expected: Ket, Selkup, Mansi,Mari, Chuvash, Saami, Finnish, Estonian...And then the IE start to appear as Latvian, Lithuanian, Russian_Kargopol though still mixed with Erzya, Mordovian... Nothing surprising, anyway.

Overall a very good run. Thanks again.

DMXX said...

^ Cosign the above in its entirety, Alberto.

On the topic of IBD segments, though not an equivalent body of evidence, the user "Kurd" at Anthrogenica has started a mini-project comparing segment sharing between users and ancient samples there. Right now, South-Central Asians and Kurds appear to be roughly equivalent in their total segment count with Kotias.

I don't know if anyone has undertaken IBD analysis involving Kotias just yet, but I agree it'd be an illuminating exercise for the reason you've outlined.

FrankN said...

@epoch: "Mallory says otherwise and considers that a problem. Read about it here"

Mallory tends to be quite generous in assigning terms to Indoeuropean. Some examples:

- *grhₐnóm "grain" needs to be taken with a grain of salt. Not neccessary an agricultural term, and "corn" vs. "acorn" demonstrates possible Paleolithic use. Various parallels in "Nostratic" and beyond (c.f. Malay garam "(grain of) salt").

- The same applies to "grinding" terms. Grinding of nuts, acorn etc. is archeologically well attested from the Paleolithic onwards.
On PIE *melh² "to grind" see Uralic *molV "to break, crumble", Altaic *móĺe "to bite, gnaw".
*p(e)is "grind" has parallels in Uralic, Altaic, Korean, Kartvelian, Dravidian and AfroAsiatic with the general sense of "tear, cut, crumble, mince, squeeze" (c.f. Germ. beissen "to bite" on IE semantic variation).
The only European language knowing *h²el "grind" seems to be Greek. Considering the intensity of Greek-Armenian-Persian language contact during antiquity, this isn't sufficient to convince me of that root's "universal PIE" character.

- *pano "millet": Millet was domesticated around 7.000 BC in China. We could have a Wanderwort here that entered IE from East Eurasia (c.f. Chin. pei "black millet", Burush. *bay "millet")̣
https://en.wikipedia.org/wiki/Millet#History

- *drhxweh² [*darǝw-] "grain": Quite some twisting by Mallory - there is Engl. tare "weed, vetch (vicus sp.)", lat. dravoca "bluegrass (lolium sp.)", OInd. dū́rvā "bent grass, devil's grass (Cynodon dactylon)", derived from PIE *der- "to cut, split, tear". Correspondences include Mong. *darki "brushwood", Kor. *tār "reed", Turk. *darig "millet". This last one seems to be the only term related to agriculture, in a language family that isn't prime suspect for being neolithic pioneers, and for a crop introduced from China. Hmm..

- *ses(i)ós "grain": The PIE case is rather weak - aside from OInd. sasa "herb, grass, corn", there is only a hypothetical Gaulish (s)asiam "rye" reconstructed from a Latin text, without parallels in living Celtic. The East Asian case is much stronger: NCauc *sū(l)sū(li) "a.k. of cereal", Turkic *sɨs "grass", Jap. *sása "small bamboo", Sino-Tibet. *chuāH "grass"́ etc. Abayev (1979) suggests Central Asian origin.

- *seh "sow": Well represented in European IE, but I miss the "Asian" parallels. Standard IE dictionnaries have Hitt. sai-/ sija "to press, imprint", as cognate, which already requires quite some fantasy to be regarded as agricultural terminology. This holds even more true for OInd sā́yaka- "intended or fitted to be discharged or hurled, missile, arrow".

- *ghel is one of those PIE roots that are good for almost everything - gold, glide, yellow/ green, gall/intestines, swallow (bird), tortoise, lip, etc. "plough" is also attested - for OIndian and Armenian, none of which I would call an European branch of IE (but Mallory may have a different understanding of 'European').

This isn't to say that there are no common PIE roots for agricultural terms - maybe there are. But identifying them needs much more in-depth analysis than what Mallory has been providing. Was his presentation peer-reviewed?

For further reference, check out the database below, and search for "sow", "ghel" etc.
http://starling.rinet.ru/cgi-bin/query.cgi?root=config&morpho=0&basename=\data\ie\piet

Chad Rohlfsen said...

Here's a K8. It's pretty stripped down. Best of all, things look good unsupervised.

https://docs.google.com/spreadsheets/d/1wuFChh1HgwTc5ydgjcLETAP8vYixEQD13gGj9enAtTA/edit?usp=sharing

FrankN said...

@Dave: Here my takes on your revised K10, relating to my previous comments on the earlier version:

1. EHG/WHG in SSA: Hasn't really improved. Mota still with 1.4% EHG (a bit down) and 0.2% WHG (unchanged). EHG/WHG in East Africans appears to slightly have gone up, while Yoruba are still shown unadmixed.

2. Unexpected Oceanian admix has essentially remained unchanged.

3. Pre-LN EHG in the West/ Caucasus: Shares appear to have increased a bit.
- Kotias 5.5% (+0.3)
- Anat_Neol I0727 5.0% (-0.1)
- Anat_Neol I0725 4.5% (+0.4)
- Hungary_EN I1499 3.9% (--)
- Hungary_EN I1498 3.4% (+0.1)
- Loschbaur 4.3% (+1.1)
- Anat_Neol I0724 3.4% (+0.3)
- LBK_EN I1550 3.1% (+0.2)

etc. Apparently no artefact. The East End R1 boys seem to have been earlier and further around than expected.

4. Pre-LN CHG in the West: Minor decrease, but the general pattern remains unchanged.
CHG averages:
- Anat_Neol 1.9% (-0.1)
- LBK_EN 1.3% (--)
- Hungary_EN 0.7% (--)

Highlights:
- Anat_Neol I0708 13.2% (+0.4)
- LBK_EN I0048 7,1% (-0.1)
- Anat_Neol I1098 7.2% (--)
- Anat_Neol I0726 5.2% (-1.1)
- Anat_Neol I0707 5.6% (--)
- LBKT_EN I0176 4.6% (-0.2)

5. Unexpected SSA-Admix in non-Iberian aDNA appears to actually have slightly increased. While several RISE samples stand out, SSA admix is also shown elsewhere, e.g.
- Anat_Neol I1102 1.9%
- BB_Germ I0806 1.7%
- Srubnaya I0431 1.7%
- Kotias 1.5%
- Anat_Neol I1098 1.4%

Interestingly, no SSA admix in Iberia_EN, and only a little bit in Iberia_MN.

Maybe you might want to recheck your SSA vs. AfroEurasian allocation. Stuff like avg. 1.5% SSA vs. 0.02% AfroEurasian in Armenian_BA just doesn't feel right - neither does 1.8% AfrEur vs. 0.4% SSA with the Spanish. I mean - the figures themselves look ok, they just should be reverted.

6. SEA admix in strange places (CW, BB etc.) seems to have virtually stayed unchanged. We probably need Central Asian/IVC aDNA to understand better what this SEA entails and where it came from.

Final question: Are there any differences between "Nogai" and "Nogay", and between "Mongola" and "Mongolian", or may they safely be united under one label?

FrankN said...

@Chad "Here's a supervised K11."
Could you enable downloading of the file. There seem to be a couple of interesting things showing up, especially when it comes to identifying African-SEA relations, which I would like to take a closer look at. We know, e.g., that Bananas reached West Africa by boat from Papua or Eastern Indonesia around 500 BC - that should have involved genetic exchange. The Mandi in PNG regularly cluster in automated lexical comparisons with the Mande languages - strange coincidence...

Chad Rohlfsen said...

I can't share any of the files, sorry. You can look for Papuan and Atayal-like admixture in the Africans listed in the K8. That is the one which I am focusing on.

James said...

Davidski:

"Most of the CHG and J in South Asia arrived there with Neolithic farmers, possibly Dravidian speakers."

Dravidians aren't the ones who are high in J2. Upper caste Brahmins are - the group that "had the highest genetic contact with the Aryan-speaking peoples" according to the Mukherjee et al. paper.

"That kills your CHG = PIE theory, because Indo-Europeans move into West and South Asia from the steppe later."

That's no problem. It could still have happened that way: Indo-European J2/CHG people migrate north, mix with steppe R1a1/EHG people, then that new group moves down into Iran/Afghanistan and then India.

"Your theory is as loopy as Renfrew's dead Anatolian hypothesis."

Anatolia's pretty close to the likely source of CHG, so I wouldn't say Renfrew's hypothesis is totally dead, and certainly not loopy. What's loopy is a Slav nationalist thinking IE languages came from the steppe when nothing else did.

Chris Davies said...

@ FrankN - "Bananas reached West Africa by boat from Papua or Eastern Indonesia around 500 BC - that should have involved genetic exchange."

Banana phytoliths dated to 500 BCE were found in Cameroon by Belgian scientists, but banana phytoliths dated to 3000 BCE have been found in Uganda since then. Are you sure that after Bananas reached East Africa from India or Southeast Asia they didn't then reach West Africa by an overland route from East Africa?

Papuans are characterised by very high frequencies of HLA-B allele B*56:01 ranging from 11.5% - 35% [highest frequencies in the world, alongside Australian Aborigines]. When I looked at the frequencies of the allele within Africa, they were almost universally low [0% - 0.8%] with two exceptions: Senegal Niokholo Mandenka [N=165] @ 2.70% and Sudan Mixed [N=200] @ 2.50%. The haplotype in the Mandenka is: A*01:01-B*56:01-C*01:01 . This haplotype is not found in Papuans, where A*01:01 is at 0%. Instead it appears to be from Sudan.

Davidski said...

James,

Thanks to ancient DNA it's now easy to model the Neolithic and Indo-European expansions as separate phenomenon for Europe and South Asia. I've done it plenty of times, and Jones et al. did it too recently for South Asia.

http://3.bp.blogspot.com/-3I3MvcUynOM/VkpnFX2QNrI/AAAAAAAADsY/QFkR7el7Bc0/s1600/Jones_Supplementary_Table_10.png

The Neolithic timeframe is simply too far back to be associated with the Proto-Indo-Europeans. This is the consensus among historical linguists, who do consider Renfrew's theory loopy and instead prefer to date the Proto-Indo-European expansion to the beginning of the Bronze Age.

J2 in India is old and diverse, and can be associated with the Neolithic expansion. So there goes that. The Vedic Aryans arrived in India only around 1700 BC from the steppe and brought R1a with them.

You're obviously of Southern European extraction, so you really want to believe Renfrew's politically correct new age take on the Indo-European expansion, and Dienekes' blathering online, because Southern Europeans have a lot of ancestry from the Near East.

But you must understand that it's all bullshit. The Kurgan steppe hypothesis looks better than ever thanks to ancient DNA, and will be set out in detail and confirmed beyond any reasonable doubt next year.

FrankN said...

@Chris Davies: "Banana phytoliths dated to 3000 BCE have been found in Uganda since then. Are you sure that after Bananas reached East Africa from India or Southeast Asia they didn't then reach West Africa by an overland route from East Africa?"

Check it out here:
http://www.pnas.org/content/108/28/11311.full

East African phytolits are questionable. They may also have belonged to Enset, a banana-related species native to Africa. Nevertheless, a number of arguments have been put forward that Bananas may already reached East Africa by around 3000 BC, including the high genetic diversity of East African Bananas, and linguistic considerations (pre-Bantu substrate, see R. Blench's study below)

http://lib-ojs3.lib.sfu.ca:8114/era/index.php/era/article/viewArticle/354

While E.African Bananas are AAA triploids, the W.African ones are AAB. The additional hybridization with the B-genome must have occured in South East Asia, as the B genome is only present in wild varieties from Southern China and the SEA mainland. The most likely scenario is that AAB triploids emerged on/near the southern Phillipines as a consequence of the Austronesian "Out of Taiwan" expansion, which brought B-genome Bananas into the insular SEA/Melanesian habitat of A-genome Bananas and early AA/AAA hybrids (believed to have originated in New Guinea).
As AAB triploids are lacking among traditional East African cultivars, and AAA among West African, there seem to have been two independent waves of transfer, from different geographical regions and most likely at different points in time. Blench, btw, is furthermore making the point that the ecology of Central Africa and the Great Lakes area (dry Savanna aka as "Steppe") isn't conducive to a cross-African plant migration, while coastal deserts, especially the Kalahari, preclude non-human-mitigated expansion along the coast.
The first written description of Bananas comes from Alexander the Great's invasion of India. This demonstrates that in the 4th century BC the Banana was still unknown in the Mediterranean, including the middle/lower Nile, while already being cultivated in West Africa. Thus, a south-Saharan trajectory (which would have had to cross the Ethiopian highland before) is equally unlikely.

FrankN said...

@Chris Davies: "This haplotype is not found in Papuans.."
Interesting information on the HLA-B allele.
Linguistically, New Guinea is the most diverse part of the world, with nearly 900 languages belonging to at least five different macrofamilies. I could imagine the linguistic diversity being mirrored in genetic diversity. Thus "is not found in Papuans" is possibly better phrased as "hasn't yet been found among the rather limited subset of Papuans sampled so far". Note that there are some genetic studies from East Indonesia that have shown high diversity, including several previously unknown yDNA and mtDNA haplotypes, among Papuan-speaking populations in West Timor, Flores and on surrounding smaller islands.

The Sudanese HLA-B allele connection is also interesting. The Mande - Mandi (PNG) lexicostatistc cluster also includes several South Sudanese languages, notably the Moru-Ma'di and Lendu groups, plus, on a separate branch, Kartvelian languages.

Kristiina said...

This is to be read with FranN’s well-founded comments!

@epoch2013 I checked Mallory’s agricultural roots and there are several that fulfil the criteria I used when picking up the words onto my list. I also noticed that most of these words have a (possible) cognate word in Finnic languages.
*(s)palǝw-/-e ‘straw, chaff’- similar meaning in Balto-Slavic, Old Indian and Latin; cf. Finnish pala ‘bit’
*keres- ‘millet’, Hittite karash emmer wheat, Germanic hirse millet, Latin cerēs bread, grain, Kalasha karasha ‘millet’; cf. Finnish kyrsä ‘piece of bread’, Erzya kše ‘bread’
*aro ‘plough’ attested in a similar meaning in Tocharian, Armenian, Old Greek, Balto-Slavic, Germanic, Latin and Celtic but not in Indo-Aryan; cf. Finnish aura ‘plough’
*yewǝ- attested in Hittite, Old Indian, Avestan, Iranian, Old Greek, Balto-Slavic; cf. Finnish jyvä ‘grain’, Mokša juv ‘husks’, Udmurt ju ‘winnowed grain’
*ad- Hittite ’ein Getreide’, Tocharian aati ’grass, Armenian hat ’grain’, Latin ador ’spelt’; cf. Finnish itu, Estonian ido ‘germ’
*dhōn- cereals, bread, attested in a similar meaning in Tocharian B tāno 'seed', Old Indian dhānā́ 'corn', North Persian dān 'korn', Lithuanian dúona bread; cf. Japanese and Ainu tane 'seed'

However, some of Mallory’s roots have a restricted distribution, so they may be local inventions or substrate words. According to Mallory and Starostin’s Tower of Babel:
*sarp- Old Greek, Balto-Slavic
*āl- ‘esculent root’ Old Indian ālu 'esculent root', Latin ālum 'Symphytum petraeum', Osk *allō 'garlic
*gel- ‘plough’, attested only in Old Indian and Armenian
*g'er[a]n- ‘grain’ attested in this meaning only in Balto-Slavic, Germanic, Latin and Celtic
*bhars ‘grain, barley’ attested only in Slavic, Germanic, Latin, Italic and Celtic
*h2elbit- ‘barley’ attested only in Greek and perhaps in East Iranian *arbasyā
*g'herǝzdh- ‘barley’ attested in Hittite, Latin and Old Greek, Baltic, Germanic, Albanian
*ses(i)ós ‘grain’ attested in Sanskrit sasyám ‘grain’, Avestan hahya- ‘providing grain’ and Hittite sesa ‘fruit’
*ak- ear of cereals, Latin agna ‘ähre’, acus ‘granne’, Greek akostǟ́ ‘gerste’, OHG agana ‘spreu’, Baltic akúotas ‘Granne’, Tocharian B āk ‘ear of grain?’

Some IE roots are related to agriculture only in Europe:
*sēy- meaning in Old Indian is missile, arrow and in Hittite festdrücken
*mol(w)ǝ- ‘grind’ is not found in Indo-Iranian and in Tocharian it means “crush, oppress” and ‘wound, damage’.

Some are dubious:
*h2ekstí- ‘awn’, NWelsh eithin ‘furze’ Lithuanian akstìs ‘spit for roasting’, Russian osti ‘awn, bristle’, Tocharian āśce ‘head’
*wers- ‘thresh’, Hittite wars 'abwischen, Old Greek érrō (weg)gehen, untergehen, Latin verrō ‘schleifen’
*meigh ‘grain’, Old Irish miiach ‘measure of grain’, Lithuanian miežis ‘barley’, Khotanese mässa ‘feld’
*drhxweh ‘grain’ Gaulish dravoca ‘darnel’, NDutch tarwe ‘wheat’, Sanskrit durva ‘panic grass’
*kāpos ‘field’, Greek kêpos ‘garden’, Roshani kāpos ‘cultivated field’
*pano- ‘millet’, Latin pānicum, Shughni piinj ‘millet’
*peis- ‘grind’
*haéireh- ‘ryegrass’, Greek aírai ‘ryegrass’, Sanskrit erakā ‘sedge’

Moreover, the picture is very complex if we take into account the similar roots in non-IE languages (Caucasian languages, Turkic and Uralic languages, Burusho language and Sumerian language and in East Eurasian languages - which were referred to by FranN - , which I have not done here as it would take too much time.

Chris Davies said...

@ FrankN - "Thus "is not found in Papuans" is possibly better phrased as "hasn't yet been found among the rather limited subset of Papuans sampled so far"."

Almost 500 individuals from 7 different Papuan populations were sampled [East New Britain Rabaul; Eastern Highlands Goroka Asaro; Karimui Plateau Pawaia; Madang; Wanigela Keapara; West Schrader Ranges Haruai; Wosera Abelam]. Total HLA A*01:01 was 0.0% in all populations. The A*01-B*56 haplotype found in the Senegalese Mandenka and Sudanese is also found at low frequencies in Mediterranean populations. Papuans and Australian Aborigines tend to have A*34-B*56 and one or two others but no A*01-. A*01 seems to be at greatest frequency and diversity in Kenya Nandi.

FrankN said...

@Chris Davies: From that list, it seems that speakers of the Marienberg languages, to which Mandi (PNG) belongs, haven't yet been sampled. The closest population sampled appears to reside some 100km away (Google maps can't find a connecting road, so the distance is my estimate). All sampled populations seem to come from the central highlands, while Marienberg languages are being spoken in the southern part of the coastal Torricelli ridge, SE of the town of Mandi. The Mandi (PNG) language is spoken in just one village, by less than 500 people.

http://glottolog.org/resource/languoid/id/mari1433

FrankN said...

@Kristina: Thank you for your compliments.

A few more remarks on Mallory's roots:
*(s)palǝw-/-e ‘straw, chaff’ Possibly a reflex of PIE *pel(ew) "skin, hide". Non-IE parallels are rare: Mandekan *gbolo, Cush. West Rift *fala, both "skin".
But when allowing for "r"-"l" sound change, the root has parallels in Uralic (e.g. Hung. bör "skin, fur", Khany pĕr "birch bark"), AfrAs *pVr(w) (Heb. parwa "fur", EChad *pir "to shell, bark", Berber *fur "bark" etc.), Dravidian (e.g. Telugu beraḍu "bark, rind, shell"), Ijoid *opara "skin" and, intriguingly, Tupi-Guarani *pir-, Guahiban *perabo "skin".
Mallory's "PIE neolithic root" may in fact be a reflex of a paleolithic word with global distribution.

*keres- ‘millet’: c.f. Kor. *kắràč "wild foxtail millet (Setaria vir.), darnel ryegrass (Lolium temul.)", Tungus-Manchu *kara(~x-) "darnel ryegrass", Mong. *karaɣu "darnel ryegrass, elymus". A Wanderwort. From East Asia, for a cereal domesticated in East Asia. I think, the PIE case is closed.

*g'herǝzdh- ‘barley’ sounds suspiciously close to *keres- ‘millet’ discussed above. Hitt. karas and Baltic *girsa should rather go there. That leaves us with Alb. drith, OGrk kritha, Lat. hordeum, Germ *girsto, united by the meaning "barley", but not via common IE sound laws. This is hardly a shared PIE root, at best parallel borrowing from a yet unknown substrate [Note, however, that Germ. *girsto has a straightforward etymology from garstig "spiky".]

*aro ‘plough’: Obviously related to PIE *er- "earth, soil, arable land" [Celt/Germ/Grk, Arm/Hitt/Toch w. sem. shift]. Parallels hereto are Altaic *ā́rV "open space", Ural *arV "grass-covered lowland", NCauc *ʔārV "plain", Basque *haran "valley", Burush. har "river bed", SDrav *ar- "rivermouth". Looks like a West Eurasian HG term.
Anyway - first archeological evidence of the plough is from Northern Germany, ca. 3400 BC. Its predecessor, the ard (sic!), is documented somewhat earler from CT. So, we are dealing with a MN invention, probably in an area that is today IE speaking. So far, so good. Basque *arhe should be borrowed.
Somehow irritating, for the absence of this PIE root in Indo-Aryan, is Burush. har, Drav. *ar- "plough", Tibet. adru "to dig", Chin. r(h)ǝ "shovel".

*ad- Tocharian aati ’grass'.: That's indeed a quite strong one. But note Turkic *ot "grass", Mong. otul "(cut) reed". Both borrowed from PIE? Or are we dealing with an ancient "steppe" term?

*dhōn- cereals, bread: Possibly, the Tan(d)ur belongs here. In that case, we would be dealing with a neolithic root (tandurs have been found in LBK settlements in Hesse), though one shared by IE with Semitic, Georgian, Dravidian and Turkic. However, there are alternative etymologies for the tanur, e.g. Akkad. tin "mud" + nuro/nura "fire". Note that the "-d-" was added in Turkic, and then spread into India
https://en.wikipedia.org/wiki/Tandoor

*āl- ‘esculent root’: c.f. Altaic *ŭĺa "hemp", Lushai (ST)hlo "a weed, a drug, medicine", Proto-Nakh *jōl (obl. ela, alo) "hay".̀ German Alraune "Mandrake (Mandragora off.)" may also belong here (on the second part see "rune", i.e. mystic, magic). Anyway, the term is more horti- than agricultural.

*bhars ‘grain, barley’: c.f. Georgian puri "bread"

James said...

Davidski,

When I said Renfrew's hypothesis is not totally dead, I thought I made it clear I was just talking about the PIE homeland around Anatolia not the Neolithic timeframe. It wasn't loopy to think IE languages might've spread with farming. It made sense. Turns out they probably spread with metal working instead, which also makes sense. Gimbutas' Kurgan stuff...now that's loopy.

"J2 in India is old and diverse, and can be associated with the Neolithic expansion. So there goes that."

You keep saying that, but I haven't seen any proof. And it still doesn't explain why J2 is the main haplogroup distinguishing the northern Indian Brahmins from the lower castes. If J2 was as old as you say, it would be more evenly distributed in the population, and if anything higher in older groups like Dravidians. So there goes that.

"You're obviously of Southern European extraction, so you really want to believe Renfrew's politically correct new age take on the Indo-European expansion, and Dienekes' blathering online, because Southern Europeans have a lot of ancestry from the Near East."

All Europeans have a lot of ancestry from the Near East - from CHG and also from the Neolithic. That won't change no matter where PIE comes from, so I don't see what it has to do with anything. And unlike you, I don't need to pin all my hopes on that one thing for my self-worth. I'm just stating what's obvious and what many others here have already stated: Everything else came out of West Asia, so it's most likely PIE did too.

"The Kurgan steppe hypothesis looks better than ever thanks to ancient DNA, and will be set out in detail and confirmed beyond any reasonable doubt next year."

I can't wait to see you proved wrong again.

Davidski said...

J2 can't be a marker of the Indo-European expansions, not into India or anywhere, because it shows long branches with deep rooted-nodes, which are characteristic of haplogroups that began expanding during the Neolithic, like G2a.

http://eurogenes.blogspot.com.au/2015/05/large-scale-recent-expansion-of.html

Typically Indo-European markers have much shallower genealogies, and in India R1a has a very shallow genealogy, dating back to around 1700 BC when the Vedic Aryans appeared there.

There may have been an expansion of a local J2 variant in India on the back of the Aryan expansion, but that's not important. Not a single J has been found in any Andronovo, Sintashta or Srubnaya burial, and if you actually care to pull your head out of your ass, there's basically a consensus among linguists that these people were the early Indo-Iranians.

When the Harappan Y-DNA is announced, it'll probably be J2, which will fit with a Neolithic expansion of CHG and J2 into India.

You and the Dienekes crowd have now been reduced to claiming that the Indo-Europeans crawled out of a valley in the Caucasus, spread their language to the steppe via their women, and then let the steppe people do all the work. It's not even funny.

Karl_K said...

"Indo-Europeans crawled out of a valley in the Caucasus, spread their language to the steppe via their women, and then let the steppe people do all the work."

This actually seems pretty reasonable. There are lots of caves in the Caucasus. If I were a pre-proto-Indo-European, that's where I would be living for sure.

FrankN said...

Kristina: I hope you are still reading this and haven't moved on to the Scythians.. I have a few question you as a Finno-Ugrian speaking linguist might be able to answer:

1. Georgian doesn't have a grammatical gender distinction, but uses prefixed qualifiers instead. One of respective pairs is khari-" "bull" vs. phuri- "cow". The concept has obviously been taken over for the North Germanic karldyr "male deer, animal". Charlemagne demonstrates sound continuity of the "Kh-"/"Ch-" initial. Russ. жеребец (zherebets) "stallion" is another example of that Georgian masculine marker being applied within IE, however, in this case on a Mordvinian root, so we might have a borrowing instead of a generic "Slavic" construction.
I suspect Georgian phuri- "cow" behind German Färse "(young) cow". Speculatively, "whore", and also German Frau "woman", might be further reflexes of phuri-.

Do you think Uralic *korya "male (animal)" might belong here as well? Are there Finnish (Finno-ugrian) terms for female animals which might be derived as phuri-constructions?
In that case we would have an interesting Kartvelian-Uralic-Germanic (-Slavic?) isogloss indicative of geographically extended language contact (when?).
[Compare also SDrav. *gūḷi "ox". Distant Kartvelian-Dravidian relation is hinted at by Jäger's automated lexical analysis placing both on the same branch, albeit at a low significance level.]

2. There seems to be a proto-Uralic *el(o) "alive, being, animal", present a.o. in Hung. lo "horse". The root seems to be present in the "elk", which would indicate an early Germanic borrowing fom Uralic.
To me, Lat. equus "horse" always looked suspiciously close to an elk encountering a cow (PIE *gwou). Is this just overinterpreting accidental sound correspondence, or could PIE h²ekwos, Lat. equus "horse" in fact reflect a proto-Uralic *el(o) applied to *gwou "cow".
[That root is of unknown origin. Comp., aside from PIE, AfrAs *kaw "bull", PNC *qɦwĕɫV- "cow, mare", ST *chu "cow, bull, Yak-cow crossbreed", Burush *chao "to milk", Tai-Kadai *gwai "water buffalo", Khoe *góɛ "cow". Apparently, and not surprising for a pastoralist term, a Wanderwort].
C.f Mari üšküž, Turkic *öküz "ox" (but also see below on these terms!).́

3. Returning to Mallory's list: I have always been wondering whether PIE *uks- "ox" wasn't somehow related to the Tibetan Yak, which has left its genetic and linguistic trace across East and South East Asia, but also seems to have provided some 10% non-Anatolian admixture into Holstein-Frisian cows (link on request, haven't bookmarked it).
Suspiciously, we have a second PIE root *ag'h- "cow, ox", present in Indoaryan and Armenian, which displays obvious semantic and phonetic parallels to *uks-, but cannot be combined with it via IE sound laws. Usually, this indicates parallel borrowing from the same, non-IE source. What's your take on this?
C.f. Jap. 'usi "cow, bull", Kartv. *usxo "sacrificial bull", AustrAs *r(i)ak "buffalo"

capra internetensis said...

@James

I have no dog in this fight, but I haven't noticed any really strong patterning for J2 in India apart from being rarer in most tribal populations. This contrasts with R1a1, which correlates quite well with being upper caste and Indo-Aryan/northern.

Uttar Pradesh Brahmins according to Zhao et al had 12% J2, while Uttar Pradesh Muslims had 15-19% J2. But this level of J2 is unremarkable, often found among non-Brahmin populations of India. In Trivedi's sample of Indo-Aryan caste members (mainly from the north and east) J2 does increase with caste status, but this pattern is not particularly significant and is not reflected in other parts of India.

Among the 1000 Genomes Indian samples, Telugus have the lowest J2 (7%) but Tamils have the highest (17%) (Punjabis from Pakistan have 22%.)

In ArunKumar's large sample of Tamil Nadu (n=1680), J2 was actually slightly more frequent overall in *hill tribes* (8%) than in Brahmins (7%), and most common in scheduled castes (19%) - middle castes had 11%. (The Dravidian upper caste sample of Trivedi likewise had 8% J2, and the middle caste 16%, but in contrast the low caste sample had only 2%.) This does not follow the pattern of R1a1, which was most common in Brahmins (42%) and least common in scheduled castes (8%) and hill tribes (4%).

Usually tribals have only 0-5% J2, but West Indian hill tribes (n=382, 11 populations from 2 studies) have average 13% J2 (0-27%) - but they also have quite high R1a1 at 21% average (6-62%).

There is no clear correlation between J2 and R1a1 at the fringes: Munda (from a huge sample) had 4% J2 and 6% R1a1, hill tribes of Bangladesh 1% of each, while Ladhakis have 2% J2 and 17% R1a1, Nepalis have 7% J2 and 21% R1a1.

I wouldn't be surprised to find that there are certain subclades of J2 associated with Brahmins and/or Indo-Aryans, but overall I think that J2 is much older in South Asia than R1a1 and comes from multiple sources, quite likely including Neolithic farmers (Dravidian or otherwise).

Arch Hades said...

So Indians have a shit ton of young/shallow R1a, but no EHG autosomally? That doesnt make much sense. Do not even the Kalash have EHG?

Davidski said...

How are these D-stats possible if there's no EHG among upper caste Indo-Aryans?

http://eurogenes.blogspot.com.au/2015/11/caucasus-hunter-gatherers-chg-and-indo.html

mickeydodds1 said...

'Everything else came out of west Asia, so it's likely that PIE did too'.

Ever since I began to study science, way, way back in high school, I was taught to never, ever to use that type of thinking, or at the very least to publish it openly, since it is the biggest and worst possible scientific-method error of thinking possible.

If we compare it to the standards of accepted etiquette, it's rather like belching openly at the dinner table.

Davidski said...

It's circular reasoning. The assumption is made that "everything" came from West Asia, so everything of note must come from West Asia.

That's the type of thinking that made a lot of people claim Sintashta were Near Eastern migrants in Eastern Europe. Clearly they were not, as ancient DNA shows that they migrated to their homeland near the Urals from further west in Europe.

It appears to be a similar mental illness to the one suffered by Afrocentrists, except in this case we have Southern Europeans claiming achievements for West Asians.

James said...

It's called Occam's razor. Learn to use it so you don't end up devoting your life to loopy theories.

And the comparison with Afrocentrists is way off. Southern Europeans and West Asians already have plenty of achievements. They don't need to claim anyone else's. The mental illness of Afrocentrists who *need* Egypt is rather like that of Slav nationalists who *need* PIE. They'll believe it's true no matter what, because it's all they got to feel special.

"http://eurogenes.blogspot.com.au/2015/11/caucasus-hunter-gatherers-chg-and-indo.html"

Too bad Jones et al's interpretation of the data agrees with me and Occam, and not with you.

Davidski said...

Occam's razor has been putting the PIE homeland on the steppe for 50 years or more. It's still doing it.

The ancient DNA evidence in the Jones et al. paper obviously backs the steppe hypothesis.

I'd say that you're in denial for personal reasons. But in all honesty, I have to say that you're really not the sharpest tool in the shed James.

Arch Hades said...

@James. Jones et al just said that CHG could be linked to the spread of Indo-European languages in South Asia, not that it's the only component. The steppe hypothesis just includes other components as well like EHG/ANE. I don't see anything in their conclusions that disregards the steppe. Btw, not everything in prehistory comes from West Asia. The domestication of the horse for example totally points to the steppe. It's unlikely the language would have been so succesful without it IMO.

André de Vasconcelos said...

David, I really like the work you put into these topics and into your blog - and I thank you for it, I've been learning a lot - but I feel your apparent hostility towards SEuros who disagree with you makes this comment section "less enjoyable" (to be PC).

mooreisbetter said...


Here is a review of all theories on why R1b came to be so common in Europe, including theories using the terminology like "mixed marriages."

http://snplogic.blogspot.com/2015/12/a-review-of-all-theories-on-why-r1b-is.html

Davidski said...

Andre,

Honestly it makes no difference to me whether I'm dealing with, say, a Southern European pushing Dienekes' agenda, or a Hindutva crackpot from India. If they insist on pushing an agenda without using any real arguments (eg. it must be so because it's Occam's razor) and propagating stupid shit here then I'll be very hostile.

mooreisbetter,

Considering the ancient DNA that is already available all we can debate now are the details of how things happened, and not the general trends. So I think you're fighting a lost cause.

Arch Hades said...

The only thing i'm a little skeptical of is the gender based assumptions regarding the geneflow exchange between EHGs and CHGs on the steppe. If it's so one sided, why havent the authors of the recent sudies from this last year said that? You know how authors will state most the North African geneflow into Sicily is paternal based, how the majority of European ancestry in Askenazi Jews is maternal based, how most sub saharan geneflow into North Africans over the past 1500 years is maternal based..etc. Yet none yet have said the CHG in Yamnaya is mostly maternal. It would be nice if the scientists endorsed this view. Because i do think it plays a role in strengthening or weakening a particular model. As David says, The idea of "migrating women" imposing their their language onto other people is pretty laughable.

capra internetensis said...

@Arch Hades

The spread of Austronesian languages is associated with East Asian mtDNA and X chromosomes but incorporated more local Y DNA.

Not because of rampaging Amazons of course, but because of matrilocality most likely. Warlike, expansionist, matrilineal Neolithic societies are not uncommon, eg the Iroquois.

James said...

This is what it says in Haak et al 2015:

"The Steppe hypothesis gains in plausibility by our discovery of a migration during the Late Neolithic from the steppe into central Europe. This migration was predicted by some proponents of the Steppe hypothesis and we have now shown (definitively) that it occurred."

And then:

"The Armenian plateau hypothesis gains in plausibility by the fact that we have discovered evidence of admixture in the ancestry of Yamnaya steppe pastoralists, including gene flow from a population of Near Eastern ancestry for which Armenians today appear to be a reasonable surrogate."

The other hypotheses *lose* in plausibility.

And finally it says:

"However, the question of what languages were spoken by the “Eastern European hunter-gatherers” and the southern, Armenian-like, ancestral population remains open."

So the genetic data actually backs *both* remaining hypotheses equally well, and that's when you need Occam's razor to decide whether to bet on Eastern Europe or West Asia.

Will PIE be from where farming and metal working and the earliest civilizations came and spread from? Or will it be from a desolate, thinly populated place on the fringes where not much else came from?

The answer is obvious to anyone who's not a biased Slavocentrist.

Davidski said...

James,

You don't know anything about Eastern Europe. You're utterly clueless on the topic.

How can you call Eastern Europe of the pre-PIE period a "thinly populated place on the fringes where not much else came from"?

Are you out of your fucking mind? Have you never heard of the ancient proto-cities in Neolithic Ukraine and surrounds?

You appear very child-like mentally. Do you know how to Google? Can you read and comprehend simple English? Try this, hopefully it's not too high brow for you.

http://thehistoryofeuropepodcast.blogspot.com.au/2012/06/incredible-khvalynsk-culture-of-lower.html

And the Haak quotes are outdated. You should at least know that now, you moron. There was no Armenian-like population, only Caucasus hunter-gatherers.

James said...

A 2015 study outdated? Are you out of *your* fucking mind?

The Armenian-like population obviously = CHG.

Davidski said...

What you still don't want to understand is that Yamnaya doesn't have Armenian ancestry. It has Caucasus hunter-gatherer ancestry, which was probably hanging around the northern slopes of the Caucasus in EASTERN EUROPE for a very long time.

And all that crap in the Haak paper about the Armenian Plateau hypothesis looking more plausible was there to make the paper look more objective and nuanced.

Already at that time they knew that Yamnaya was not part Armenian. Note the bolded parts in the text you quoted.

The Armenian plateau hypothesis gains in plausibility by the fact that we have discovered evidence of admixture in the ancestry of Yamnaya steppe pastoralists, including gene flow from a population of Near Eastern ancestry for which Armenians today appear to be a reasonable surrogate.

So just barely reasonable at that time, but they knew the CHG genomes were coming.

Get over it. The Armenian Plateau hypothesis isn't that good or even interesting. And it even fails your own Occam's razor test, because there was fuck all in Armenia and surrounds at the time compared to the ancient cities and temples in Ukraine.

http://www.sci-news.com/archaeology/science-temple-trypillian-culture-nebelivka-ukraine-02223.html

Arch Hades said...

David, Is your absolute objection to the Armenian Plateau hypothesis because modern Armenians today are not identical to CHGs? Isn't that kind of like someone objecting to the Steppe hypothesis because modern Southern Russians aren't identical to Samaran EHGs or even Yamnaya for that matter? And have different ancestral strains at different degrees? Not very relevant I'd say, at least when talking about homelands.

The Armenian Highlands are only a few hundred miles away from where Kotias and Satsurblia [CHGs] were found in Georgia, it's unlikely the people there at the time will be significantly different IMO.

Anyway I think we need more aDNA, now that we have West Anatolia and the North Caucasus, we need other areas further south like the Levant/South Caucasus/Mesopotamia, Iraq, Iran etc. I want to know the ultimate origin for both of these ancestral strains [EEF and CHG]. Good luck getting Western scientists to explore this fanatical Islamic part of the world though. Yikes

Davidski said...

Caucasus hunter-gatherers aren't from the Armenian Plateau, they're from a refuge near the western coast of the Black Sea, which lies very close to the steppe.

If you believe that during the 3-4th Millennium BC Caucasus hunter-gatherers speaking Proto-Indo-European migrated in a massive wave to the steppe, then all the best with that.

Considering that we have ancient DNA from Middle Bronze Age Armenians, and they're nowhere near being pure Caucasus hunter-gatherers, I don't believe it.

You'd have to show me a Caucasus hunter-gatherer genome from 4th Millennium BC Armenia buried in a way that suggests something proto-Kurgan about him for me to believe it.

Otherwise, I'm just going to keep believing that the Proto-Indo-Europeans formed from the Khvalynsk and Srdeny Stog Cultures on the steppe, and acquired their Caucasus hunter-gatherer admixture via both hostile and friendly contacts with various groups from the North Caucasus.

See, this fits the DNA evidence we have, and also the linguistic and archaeological evidence, which show the pre-Proto-Indo-European language and people to be of North Eurasian origin, but with strong influences from the Caucasus.

James' "Occams razor everything comes from the Near East and nothing from Eastern Europe" theory simply looks like the ramblings of a twisted mind to me.

James said...

It's obvious Haak and Jones are talking about the *exact same* West Asian ancestral component, just giving it different names.

Haak:

"The Yamnaya differ from the EHG by sharing fewer alleles with MA1 (|Z|=6.7) suggesting a dilution of ANE ancestry between 5,000-3,000 BCE on the European steppe. This was likely due to admixture of EHG with a population related to present-day Near Easterners, as the most negative f3-statistic in the Yamnaya (giving unambiguous evidence of admixture) is observed when we model them as a mixture of EHG and present-day Near Eastern populations like Armenians (Z = -6.3; SI7)."

Jones:

"CHG genomes significantly contributed to the Yamnaya steppe herders who migrated into Europe ~3,000 BC, supporting a formative Caucasus influence on this important Early Bronze age culture. . . . CHG ancestry in these groups is supported by ADMIXTURE analysis (Fig. 1b) and admixture f3-statistics14,25 (Fig. 5), which best describe the Yamnaya as a mix of CHG and Eastern European hunter-gatherers."

So Armenian-like = CHG. And based on the data, the Armenian plateau hypothesis is just as plausible as the Steppe hypothesis. But *more* plausible based on everything else.

Davidski said...

The quotes from Haak are now irrelevant, and nothing in the Jones paper supports the Armenian Plateau hypothesis.

Like I just said, Caucasus hunter-gatherers aren't from the Armenian Plateau. They're from a refuge in western Georgia separated by mountains from the Armenian Plateau.

Davidski said...

James, actually, this quote from Haak is relevant.

A negative statistic for both Armenians and Yamnaya with each other as a reference population may suggest that a third (unsampled) population admixed into both the Yamnaya and to Armenians.

This is what the Caucasus hunter-gatherer data supports.

See, there was no migration from Armenia to the steppe. It's just that Caucasus hunter-gatherer DNA diffused both onto the steppe and into Armenia.

James said...

No, Haak is not irrelevant. He used Armenians as a surrogate for that "third" population because they're very closely related to it and from the same geographic area.

CHGs aren't from near the steppe or eastern Europe, and they're not related to EHGs. According to Jones, their origins are *south* of the Caucasus, and they're related to Neolithic early farmers (West Asians).

"Given their geographic origin, it seems likely that CHG and EF are the descendants of early colonists from Africa who stopped south of the Caucasus, in an area stretching south to the Levant and possibly east towards Central and South Asia."

So the migration into the Caucasus and then the steppe came from the Armenia/Anatolia/Levant/Iran (West Asia) region, and PIE could have come with it.

Davidski said...

So the migration into the Caucasus and then the steppe came from the Armenia/Anatolia/Levant/Iran (West Asia) region, and PIE could have come with it.

Surely you can't be this dumb.

CHG are native to Upper Paleolithic Satsurblia, western Georgia. And Yamnaya have CHG admixture, not Near Eastern admixture. Right?

So answer me this; if Yamnaya are a mixture of EHG and CHG, then how are you seeing a migration to the steppe from anywhere south of Satsurblia?

Arch Hades said...

So David you think CHG ancestry was very present in Neolithic North India and Iran bringing all that J2 as you say in your earlier comments, but not just a hop, skip, and a jump away in Armenia? That doesnt make sense. But you raise a good point, if we are talking about migration say after 6,000 BC or so then there is no need for the CHG ancestry found in the steppe to come South of Geogia. The quote James is pasting is refering to a very OLD common relationship between EF and CHG. This goes back to like 20,000 BC.

Davidski said...

I never said that CHG wasn't present in Armenia during the Neolithic. What I said was that there's no evidence that the CHG on the steppe comes from Early Bronze Age Armenia.

The discovery of CHG in Upper Paleolithic/Mesolithic Satsurblia and its presence already in the North Caspian Khvalynsk remains does not help the Armenian hypothesis.

Rob said...

Right, so this CHG didn't have Basal farmer stuff, but was admixed with some form of ANE. So it must have come from Central Asia; and it wasn't in EE before 4000 BC.

And Kurgan-type burials *have*bbeen found south of the caucasus; as early as 3800 BC.

Old stuff rehashed from 2 years ago.
We now just have hard proof

Davidski said...

CHG was in the western Caucasus right near the steppe 12,000 years ago.

There's no reason to believe it entered the steppe from Central Asia or Armenia, and there's no reason to believe it brought R1a or R1b to the steppe.

Like I said, the discovery of CHG in Upper Paleolithic Satsurblia doesn't help any hypothesis except the steppe hypothesis.

Davidski said...

We do have evidence that CHG was mixing into Khvalynsk and Sredny Stog steppe populations from the onset of the Eneolithic via mixed marriages with groups from the North Caucasus.

Again, this doesn't support any hypothesis except the steppe hypothesis.

What evidence do you have that there was an invasion of the steppe from Armenia at any stage?

Rob said...
This comment has been removed by the author.
James said...

Yamnaya *do* have Near Eastern admixture, no matter how you look at it. CHG represents a *southern* population that's related to early farmers and basal Eurasians from a large area across the Near East.

Just because some of its northernmost members were "hanging around" sort of near the steppe doesn't make it a steppe population ("teal" and "West_Asian" components, remember?).

There's also no evidence the migration into the steppe started from that exact spot in western Georgia where the UP remains were found. It's much more likely it came from farther south with the spread of metal working technology.

http://4.bp.blogspot.com/-Huw43FBUPdw/UAxcBjHKs1I/AAAAAAAAFEg/ppbvEuHJyf0/s1600/metallurgy.jpg

Either way, if PIE is found to come from CHG rather than EHG, then your Slavocentric Indo-European fantasy is dead.

Davidski said...

@James,

It's much more likely it came from farther south with the spread of metal working technology.

http://4.bp.blogspot.com/-Huw43FBUPdw/UAxcBjHKs1I/AAAAAAAAFEg/ppbvEuHJyf0


The early steppe groups don't have admixture from this region on the map.

PIE didn't come from the Near East or the Caucasus. You're delusional.

@Rob,

Yes, the mixed marriages increased as mobility and wealth increased on the steppe during the late PIE period.

Either that, or there was an invasion of the steppe by CHG PIE women, because there's not a single Y-HG J in any of the Bronze Age elite Kurgans. Haha.

Rob said...
This comment has been removed by the author.
jv said...

Agree, yes the Sredny Stog ( pre Yamnaya) could have mixed with the CHG. My MtDNA H6 expanded with the Pontic Steppe Yamnaya Culture but my ancient Grandmother may have come up through the Caucasus and expanded with YDNA R1b and R1a. MtDNA H6a1b has been found in a male Samara Russia male, R1b Yannaya Culture. Her line H6a1a is associated with the Corded Ware Culture. Also, H6a1a was obtained from a Srubnaya Culture site in Russia. I'm wondering if H6a1 may have had origins in the Maykop culture VIA the Iranian Plateau as H6a1a has matches in current Ossetian populations and other Iranian populations. jv

jv said...

Also, as I reread some of the comments, Samara Russian DNA is VERY similar to German Corded Ware. MtDNA H6a1b is from Samara. MtDNA H6a1a from Esperstedt Germany. My ggg Grandmother was from Northern Bavaria. Our MtDNA is H6a1a2b1. jv

Kristiina said...

Frank, true, I have not been reading that post any more, so I had not noticed your questions.

Kor-root is found in all Uralic languages except for Mari, Mordvin and Udmurt: all Finnic languages, e.g. Finnish ‘koira’, dog, ‘koiras’, male animal; Komi ‘ki̮r’, male dog; Mansi ‘kēr’, Khanty ‘kar’, male animal, Hungarian ‘here, drone (bee), testicle; Nenets ‘xora’, Taugi ‘kuru’, Selkup ‘qor’, Kamas ‘kora’, male animal. However, I see that on the IE side, there are the following words: Armenia ‘aru’, ‘vord͡z’, male animal; Persian ‘xorus’, male animal; Breton ‘gour’, Welsh ‘gwryw’, male animal; Greek ‘arren’, male animal; Lithuanian ‘vyras’, Latin ‘vir’, Spanish ‘varón’, Sanskrit ‘vṛṣa’ man. On the other side, there are also the following words: Basque ‘hor’, dog, ‘ar’, ‘orots’, male animal; ingush ‘ärh’, Chechen ‘ēra’, not castrated; Hunza burusho ‘hir’, man. To my surprise, I notice that a construction for the Proto-Chinese word ‘dog’ is ‘kkhwirʔ’. On the Turkic side, we have the Mongol word ‘er’, male animal and pan-Turkic ar/ir/er for man, and the Turkic word ‘erkek’, male. So, apart from the root ‘kVr’, we have the root ‘Vr’, which is also attested in Uralic languages: Finnish ‘uros’, male, N Saami ‘varres’, male; Udmurt ‘vorgoron’, man; Komi ‘verös’, man Hungarian ‘úr’, lord (cfr. Turkic ‘urï’ young man). Moreover, the Urartuan word for man seems to be ‘ayr=s’ and there is even the Sumerian word ‘ur’, young man, dog. Therefore, this root is incredibly widespread in Western Eurasia, and to me it looks like it has been going around in several waves in many areas.

I could not find any root for ‘woman’ or ‘female’ resembling ‘pur’ in Uralic languages. However, in the Tower of Babel Georgian root is grouped with the following IE roots:
Old Indian: pŕ̥thuka- m. `boy, young of any animal'
Armenian: orth, -u `Kalb des Rindes od. Hirsches'
Old Greek: póri-s, -ios f., pórti-s, -ios f. `Kalb, Färse'
Here, we can find a similar Finnic root: Karelian ‘petra’, deer, reindeer; Estonian ‘põder’ reindeer; Finnish ‘peura’, deer, ‘poro’, reindeer; North Saami ‘bärtoš’, trap for hunting deer. The widespread Uralic root for young deer is ‘*poča’.

The corrispondences I coul find for ‘elk’ are the following: Mansi ‘low’, Khanty ‘loɣ’, horse, KhantyO ‘kɒlȧŋ’, reindeer; Tuvan ‘elik’, Bashkir ‘ilek’, wild goat, Evenk ‘elkēn’, deer; Ingush ‘ghala’ deer; Koryak ?əlwan, Chukchi ‘əlwelu’, deer; Chinese ‘(mǔ)lù’ deer; Nuosu (Tibeto-Burman) ‘lɯ³³’ cow. On the IE side, a similar root is *elen- , e.g. Slavic *elenь, Irish ‘eile’, elk. Therefore, I am not sure if Hungarian ‘lo’ can be derived from the Uralic root ‘to live’ and if the Germanic form could be derived from a Uralic source.

Kristiina said...

Frank, I agree that roots similar to PIE *uks- are also abundant. Apart from what you listen, there are also Hungarian ‘ökör’, ox; Khanty ūgəs pl uksət, bull; Udmurt ‘oš’, Komi ‘öš’, bull, ox; Mongol ‘üxer’, ox; Abkhaz ‘ážw’, cow Nivkh ‘eӽaŋ’ cow. The Khanty word reminds me of Finnic ‘ukko’, old man, which is usually considered as derived from ‘uros’, male.

It is not an easy task to put all these words in a nice 'tree' that corrisponds to their lines of expansion.

FrankN said...

Kristina, thx for the answer!
"It is not an easy task to put all these words in a nice 'tree' that corrisponds to their lines of expansion."
Indeed! In all likelyhood, a network should be more appropriate than a tree. Whether genetic relation, Sprachbund or whatever else - evidence of long and quite intensive trans-Eurasian exchange, via the Steppe, but also forest and tundra (AFAIK, elks aren't really Steppe animals...). The Altaic roots of the "mare", OHG marcos (as in Marcomanni) are present to you, I assume.

"On the IE side, a similar root is *elen- , e.g. Slavic *elenь, Irish ‘eile’, elk. Therefore, I am not sure if Hungarian ‘lo’ can be derived from the Uralic root ‘to live’ and if the Germanic form could be derived from a Uralic source."
OHG Elen is an alternative word for the elk, which is one of the reasons why the Elk is so suspicious of being an Uralic (Tuvan, Bashkir, Evenk) borrowing. How it entered Germanic, I don't know (Skyths, Huns?).
Chinese ‘(mǔ)lù’ deer; Nuosu (Tibeto-Burman) ‘lɯ³³’ cow, Mansi ‘low’ elk, Hung. lo "horse" seems to point at an ancient meaning of "animated being", applied to each culture's predominant hunting game / domesticated animal. In Germanic, the relation "to be"->"being" (c.f. Germ. "[Lebe-]Wesen") is quite straight-forward. Finn. elo "to live", elain "animal" seems to represent the same relation (and is actually providing a nice potential source for PIE *elen-). As such, I do not fully understand your hesitation to see Hung. lo "horse" etc. as being derived from PUR *elä-.

On the *Vr- root, let me add Dutch Varken, LowGerm Barch "boar", possibly also the Aurochs. Osset. wyrs, dial. urs, MHG urs/ors "stallion" [Ross] should belong here as well. There are furthermore a couple of Sanskrit constructions with preceding *var- that denote male animals.

When it comes to *kVr- vs. *Vr-, it is tempting to think about contamination by Old Egyptian Ka "bull". In that case, it should have taken place somewhere between Egypt and Sumer, some time during the 3rd mill BC. However, c.f. for "man": Bambara ke, Beng (E. Mande) go, Mundari kora, Roviana (AN) koreo, Quechua kari(!). Ket ket "(Hu-)man" is another possible contamination source.
https://en.wikipedia.org/wiki/Bull_(ka_hieroglyph)

ur/wur/yur etc., btw, seems a pretty widespread term for "person" (male?) in NC languages. Maybe we should start to take yDNA E admix in SE Europe serious also from a linguistic perspective...

In any case, it is interesting to see so many IE cases of gender prefixing in a language family supposed to work by gender suffixing. Actually, there are a couple more of such instances: Aside from the Englisch he/she-goat, and fa-ther/mo-ther/bro-ther, we have lat. ho-mo/fe-mina, which could actually reflect a khori-phuri construction. The *por- "mother, infant" construction comprises, aside from your examples, also Lat. porcus, Germ Ferkel "piglet", and Lat. puer "boy" (note the *Vr- root here). Particularly intriguing is Lat. caballus "stallion" in relation to Udmurt val "horse".

As we are often dealing with agricultural terms, one might think about EHG substrate. However, the pattern extends to wild animals and horses, domesticated after the EN. Gender prefixing seems most common in the Caucasus, including now often fossilised nominal class systems in NWC, so I assume the Caucasus to be the source of the atypical gender-prefixed terms found in IE.

«Oldest ‹Older   1 – 200 of 204   Newer› Newest»