search this blog

Wednesday, February 13, 2019

Ancient Caucasus open analysis and discussion


The following samples from the recent Wang et al. paper on the genetic prehistory of the Caucasus are now in the Global25 datasheets:

Catacomb MK3003
Catacomb RK4001
Catacomb RK4002
Catacomb SA6003
Darkveti-Meshoko I1722
Darkveti-Meshoko I2055
Darkveti-Meshoko I2056
Kubano-Tersk BU2001
Kubano-Tersk GW1001
Kubano-Tersk LYG001
Kubano-Tersk MK5009
Kubano-Tersk PG2002
Kubano-Tersk RK1003
Kubano-Tersk_Late KBD001
Kubano-Tersk_Late KBD002
Kura-Araxes_Kaps ARM001
Kura-Araxes_Kaps ARM002-003
Kura-Araxes_Velikent VEK007-009
Lola NV3001
Maykop OSS001
Maykop_Late MK5001
Maykop_Late MK5004
Maykop_Late SIJ001
Maykop_Late SIJ002
Maykop_Late SIJ003
Maykop_Novosvobodnaya I6266
Maykop_Novosvobodnaya I6267
Maykop_Novosvobodnaya I6268
Maykop_Novosvobodnaya I6272
North_Caucasus_MBA KDC001
North_Caucasus_MBA KDC002
Progress_Eneolithic PG2001
Progress_Eneolithic PG2004
Steppe_Maykop AY2001
Steppe_Maykop AY2003
Steppe_Maykop SA6001
Steppe_Maykop SA6004
Steppe_Maykop_o IV3002
Steppe_Maykop_o SA6013
Vonyuchka_Eneolithic VJ1001
Yamnaya_Caucasus RK1001
Yamnaya_Caucasus RK1007
Yamnaya_Caucasus SA6010
Yamnaya_Caucasus ZO2002

A lot of people don't seem to be aware of this, but the links are always the same for all of the datasheets, even after major updates:

Global25 datasheet ancient scaled

Global25 pop averages ancient scaled

Global25 datasheet ancient

Global25 pop averages ancient

Feel free to analyze the data in any way you wish and share your findings in the comments. Did the authors miss anything?

See also...

Big deal of 2018: Yamnaya not related to Maykop

170 comments:

Arza said...

Distances of all new samples to 30 nearest individuals (scaled spreadsheet):
https://i.postimg.cc/LRQvBfVz/wang-dist.png

Off-topic:
Formation of the Indo-European branches in the light of the Archaeogenetic Revolution
John Koch

https://www.academia.edu/38336128/Formation_of_the_Indo-European_branches_in_the_light_of_the_Archaeogenetic_Revolution?source=swp_share

Anonymous said...

@Davidski In your opinion, was Y DNA J (and its CHG) really contained behind the Caucasus for the entire Mesolithic and Neolithic? That seems unlikely to me. Should we not expect to see lots of Y DNA J in the Steppe SOUTH of Stalingrad during, say, the 7th millennium BCE? And would it be unfair to associate that hypothetical presence with e.g. Rakushechny Yar (which has extremely similar pottery to Shulaveri-Shomu despite predating the latter - that could be explained by the 8.2 ky event, where the North Caucasus was depopulated and people seem to have migrated Southwards (to probably form S-S)).

Perhaps that is the population with which e.g. Steppe Eneolithic acquired its heavy CHG. Alternatively there's the Lower Volga Eneolithic (Orlovka horizon), which has links to Central Asia (Kelteminar) as part of the recolonisation of the PC Steppe after the 8.2ky event.

It's all here: https://adnaera.com/2019/01/11/how-did-chg-get-into-steppe_emba-part-2-the-pottery-neolithic/

I can see both being true for the introduction of CHG to the Steppe: with the recolonisation providing most of the CHG to e.g. Yamnaya, and the mixing of male EHG folk with female CHG remnants from the 8.2 ky event providing the additional CHG seen in the Eneolithic Steppe samples.

Ric Hern said...

@ Euxeinos Pontos

There were already CHG ancestry in Villabruna 12 000 BCE....In other words there were CHG people somewhere between the Ural Mountains and the Caucasus prior to that date. Most probably near the Lower Don from +-17 000 years ago.

Anonymous said...

@Ric Hern Really? Well now that is interesting. It changes nothing about what I said as presumably it would be minuscule, but that would imply that the original bearers of Y DNA R had CHG-like ancestry from the beginning (perhaps it originated in Central Asia rather than Siberia?)

Anyway, it changes nothing about what I said. So someone please comment on it.

Davidski said...

@Euxeinos Pontos

I'm not quite sure what you're asking me, because I'm aware of the fact that Y-haplogroup J is found in Eastern European Hunter-Gatherers (EHG) from what is now Karelia in far northern Russia.

These samples also show some Caucasus Hunter-Gatherer (CHG) ancestry, so it's definitely not an error.

But the data from the Pontic-Caspian steppe clearly point to sex-biased mating between EHG and CHG there, which resulted in R1-rich and CHG-rich Eneolithic populations in that part of Eastern Europe. Sex-biased mating between forager groups is almost standard practice, so this shouldn't be surprising.

And the explanation that I've seen that the high frequencies of R1 haplogroups in these steppe, as well as many steppe-derived, groups are due to founder effects on the Eneolithic steppe doesn't ring true with me, because there's already a wide variety of R1 and Q lineages in the few male samples from the Eneolithic steppe. Another potential explanation is that southern lineages, like J and G, were selected against in prehistoric (but not historic) steppe populations, but that's just silly.

So there was definitely sex biased mating between EHG and CHG on the steppe, and that's the primary reason why you don't see any Y-haplogroup J in Sredny Stog, Khvalynsk, Progress Eneolithic, Yamnaya, Corded Ware, Poltavka, Catacomb, Sintashta, Srubnaya, Kurgan Bell Beakers, etc.

It might eventually show up in one or more of these populations as a very minor marker, but that won't change the picture, which is already very clear.

By the way, keep in mind that this is a very specific analysis and discussion thread, and not a general discussion about topics that have already been well covered on this blog. So please stay on topic.

Anonymous said...

I meant is the archaeological context I mentioned plausible for an origin of CHG in the Eneolithic Steppe samples and in Yamnaya (if we had Orlovka samples, it would clear up things A LOT - I wonder why we don't have anything from the Lower Volga before Yamnaya...)

So for example, would those Iranian Mesolithic samples fit better for Yamnaya than Eneolithic Steppe, and would CHG proper fit better for Eneolithic Steppe than Yamnaya? Because that's what I'd predict - it's perfectly on topic.

Davidski said...

@Euxeinos Pontos

So for example, would those Iranian Mesolithic samples fit better for Yamnaya than Eneolithic Steppe, and would CHG proper fit better for Eneolithic Steppe than Yamnaya? Because that's what I'd predict - it's perfectly on topic.

Well, I can definitely tell you that Iran_Hotu and Iran_N aren't relevant to the southern ancestry in Eneolithic steppe and Yamnaya. And neither are those two BMAC samples from Gonur which are probably Kelteminar migrants, because they're a mix of West_Siberia_N and Iran_N.

The southern ancestry in Eneolithic steppe comes from a population very similar to CHG, but not exactly like it, which isn't surprising because we're probably dealing here with foragers who lived north of the Caucasus, not south of it.

And obviously Yamnaya is a mix of Eneolithic steppe, Middle Neolithic European farmers, and foragers from various parts of the steppe. I'm getting a very good model with Eneolithic steppe, Ukraine_N and GAC_Ukraine.

Ric Hern said...

@ Euxeinos Pontos

As far as I can remember there are archaeological evidence of spread from the Imereti region to Kammenaya Balka during the Upper Paleolithic. So CHG most probably not from Siberia. Like I said. Most likely between the Caucasus and the Urals somewhere along the Upper Volga and Lower Don.

Davidski said...

@Arza

Distances of all new samples to 30 nearest individuals (scaled spreadsheet): https://i.postimg.cc/LRQvBfVz/wang-dist.png

Nice quick validation of the data there.

Ric Hern said...

@ Euxeinos Pontos

Last comment from me on this. R1 most probably met CHG like people in that area that I mentioned while on their way from Siberia via the Southern Urals and down the Volga to the Lower Don. The ancestors of Villabruna and Yamnaya probably split at the Lower Don region with one migrating West to and the other South towards the Caucasus Piedmont area.

Drago said...

Some initial runs;

Yamnaya_Samara:I0444
Samara_Eneolithic:I0122 50.6%
KcEnlStp 38.9%
Trypillia 8%
Maykop_Late:MK5004 2.5%

Yamnaya_Caucasus:ZO2002
KcEnlStp 59.3%
Samara_Eneolithic:I0122 28.6%
Protoboleraz_LCA 10.6%
Maykop_Novosvobodnaya:I6266 1.5%

Yamnaya_Ukraine_outlier
Maykop_Late:MK5004 48.1%
Samara_Eneolithic:I0122 22.9%
KcEnlStp 19.2%
Trypillia 9.8%

Yep, good call Davidski.

Aram said...

FYI

The "North Caucasus" label in the paper is somewhat misleading, because it refers to different things.
North Caucasus as a geographic region and to the North Caucasian archaeologic culture. The other name for this culture is a Kubano Tersk culture. This is a EMB culture from ~3000-1500 BC period. It was known to have a more northern origin and now this is genetically confirmed. It is a simple offshoot of Yamna. With the same y dna.It is a steppic culture and has no presence in mountains.

The North Caucasus MBA on the other side is a geographic designation and it is not related to that Kubano Tersk culture.

Davidski said...

@Diego

I removed your comment.

The reason I did that is because it came out weird (seems like you tried posting an image directly, please don't do that).

Also, there's a thread dedicated to Villabruna at the link below. You can post about its CHG, or lack of, there.

Villabruna cluster =/= Near Eastern migrants

Matt said...

Thanks Davidski.

All, first thing I'll do is West Eurasian PCA reprocess of G25 data:

https://imgur.com/a/AlV0ibo

- Steppe_Maykop seem to overlay Sintashta_MLBA_o1 to some extent but not Sintashta_MLBA_o3. They're roughly on a cline to Dali_EBA, I guess, but Dali_EBA is so rich in West_Siberia_N its hard to tell this from just being on a cline to West_Siberia_N.

I'd judge they're more on a cline from Piedmont Eneo+Caucasus Eneo to somewhere between EHG and West_Siberia_N and they look like they are much less fitting as between Iran_N and West_Siberia_N than Dali_EBA are, though (the offset towards CHG and WHG looks too high in the relevant dimensionality, and Botai/West_Siberia_N too low).

Steppe_Maykop probably represent the least divergent of a set of western steppe populations, who were not quite purely paleolithic survivors but close to, which were overlian by expansions of Steppe_EBA toolkit and people, and absorbed on a low level.

- PG2004 is indeed more EHG/HG than the other two (PG2001 and VJ1001), at a clip of about 20-25% more ancestry. May make more sense to use the other two as a "baseline"?

Use the PG2001+VJ1001 as the "southern steppe" baseline, Samara_HG as the "northern steppe" baseline, then understand Khvalynsk (described as where southern and northern steppe people met) and PG2004 as intermediate to them?

Though Khvalysnk is more proximate for Yamnaya (a good argument for using them), the heterogenity and low sample count and intermediacy between these earlier extremes would seem to make it harder to interpret. To me anyway!

The "southern" Khvalysnk sample IO434, if they were on G25, would probably be almost exactly close to PG2004. The ADMIXTURE results for IO434 in Wang 2018 are certainly closer to PG2004 than they are to either IO0433 or IO122 - https://i.imgur.com/xw7woja.png.

- It does look like all the Caucasus samples here are "more CHG" than the previous Kura_Araxes samples, especially Darkveti-Meshoko. This is quite hard to delineate on the above metapca, because it is quite subtle and small and not so obvious in the lower order reprocessed PCA.

I've only labelled out Darveti-Meshoko in the applicable graphic, because the overlap is too close for the others, though they are subtly a bit beyond the Kura_Araxes Armenia_EBA samples towards CHG. The amounts are small compared to the most differentiated ancients, but probably quite noticeable compared to much closer recent populations.

Largely seems like it fits with paper to me (PG2004 more EHG/ANE, Steppe_Maykop clinal from southern steppe+Caucasus to West Siberia N, Caucasus populations more CHG than previous Kura-Araxes Armenia EBA samples, but subtly so).

Apologies for any lack of clarity in the graphics! This is kind of quick and dirty.

Gaska said...

@Davidski

I have not previously mentioned Villabruna. I was just asking about the claim that he had CHG.

I was not sending any image, I do not know why that space came out in the post.

By the way and although it is out of topic, affects what we have been talking about in other threads, I am referring to the last archaeological paper (one month ago) in the regions of Alto Duero and Galicia that has proven in many sites of the Atlantic coast of Iberia the coincidence of campaniform styles (AOC, MHV, Linear, Ciempozuelos), in the second quarter of the III Millennium (2,750-2,500 AC). The Portuguese and Spanish archaeologists put the Iberian migrations back on the table, doubting Olalde's assertions. It may be interesting for you because it affects the possible origin of AOC in the Neherlands (and SGC).If you are interested I can send you the link, or maybe you already know it.

Davidski said...

@Matt et al.

Here's a challenge for those exploring the data.

Can you make sense now where Anatolia_ChL and Armenia_ChL fit into the scheme of things?

And how does Darkveti-Meshoko relate to later samples in the Caucasus and Near East?

Matt said...

Another metaPCA with PAST3, using disregard groups mode to try and bring out more distinctions in Caucasus related a different way: https://i.imgur.com/9f30AoM.png

@Davidski, good point. This may clarify whether the Anatolia_Chl particularly fit on a cline involving Seh_Gabi_Chl or Darkveti_Meshoko better? Probably need a lot of nMonte testing. It'll be hard to sort out though.

Davidski said...

Here's a new qpAdm model for the Sredny Stog R1a-M417 sample. No surprises.

Sredny Stog I6561 - qpAdm Caucasus data

Matt said...

Placing new samples on NJ trees - https://imgur.com/a/inBaIJW

Matt said...

Generating a set of euclidean distance from G25, then using to see which populations are most offset between different samples:

https://imgur.com/a/3aiJ0Eq

From this I would guess that Darkveti-Meshoko is about 70:30 or 85:15 between CHG to Anatolian proper (based on the loadings), considering only those as components... but probably has some separate drift and/or admixture, so is less a CHG proxy than would appear from that consideration.

Robustly then the later populations certainly have at least both Levantine and Anatolian components coming in. Looks like particular a Levantine/Mesopotamian flow as related to Levant_Chl? (Since later Levant+Arabian Peninsula pops seem most close relative to prediction/correlation of CHG+Barcin). Relating to Davidski's ideas about Uruk expansion impacting NE?

Davidski said...

@Matt

What's your take on the substructures within the Kura-Araxes pops, especially the differences between the samples from Kaps and the later ones from Kalavan?

Samuel Andrews said...

Have the positions in the original G25 PCA changed because of the new Uralic (from months ago) & Steppe samples?

Davidski said...

@Samuel Andrews

Have the positions in the original G25 PCA changed because of the new Uralic (from months ago) & Steppe samples?

No, for one, because those samples fit perfectly into the global diversity captured by the analysis.

Samuel Andrews said...

But CHad said Sredny Stog is Caucasus_Eneolithic+UkraineHG. Does this mean Chad is wrong! Uh!

Davidski said...

I've tried modeling the Eneolithic steppe samples, including from Sredny Stog, Progress, Khvalynsk and Vonyuchka, with Darkveti-Meshoko (Eneolithic Caucasus) and only Sredny Stog might have some very minor ancestry of this type.

Also, nothing from Iran or Turan is relevant to the Eneolithic steppe samples. Not Ganj_Dareh_N, Iran_Hotu or whatever.

We picked that ages go based on mtDNA. Well done!

Samuel Andrews said...

Georgian, Abhkasian, probably several other modern Caucasians are continuation of Caucasus Chalcolithic (maybe Neolithic).

IMO, there's for sure direct Caucasus ancestry in the whole Middle East. But may not be "recent" because it already shows up in Iran & Anatolia in 4500 BC.

Kura Axes, Iran Chalcolithic, Armenia EBA are very mixed populations. A population mixed between Caucasus, Iran_N, Anatolia_N, Levant_N could be the secret successful Chalcolithic Y DNA J group people look for that spread across Middle East & into Balkans.

Samuel Andrews said...

@All

The matters of geography was a good blog post. Its main points should be emphasized many times. Most importantly Caucasus mountains served as genetic barrier between Europe & Asia.

Steppe_Eneolithic & Caucasus_Eneolithic were separated by the Caucasus. Steppe_Eneolithic led to rise of Yamnaya-like groups who became fundamental to European ancestry.

Caucasus_Eneolithic-like groups in unknown ways contributed to whole of the Middle East. They have survived in almost unadmixed form up until today with little European admixture.

The Caucasus separation, also really important shows there wasn't enough exchange between north & south for IE languages to have originated south of the Caucasus. Yamnaya really was a home-grown European population native to land North of the Caucasus & West of Ural mountains.

A said...

Which modern population are Yamnaya and Afanasievo closest to genetically?

Samuel Andrews said...

@Philippe,

West Eurasia Index.
https://docs.google.com/spreadsheets/d/1LPWAEC3dbAEDu8aBAAcxIOa5CQjuflt0f0cvhCpZ_ME/edit#gid=0

Balts are technically closest because they have ~45% Yamnaya ancestry & lots of hunter gatherer ancestry.

Gabriel said...

@Samuel Andrews

Wasn’t there steppe incursions into the Caucasus? I think that speaks for itself.

George Okromchedlishvili said...

I still think that the real barrier was the steppe and not mountains themselves.

Chad said...

LOL. Not wrong. I'll show when I merge the set.

Slumbery said...

@Matt
"Steppe_Maykop seem to overlay Sintashta_MLBA_o1 to some extent but not Sintashta_MLBA_o3"

Based on my eralier nMonte runs there are two mayor differences between o1 and o3.
Sintashta_MLBA_o3 accepts some actual Sintashta (and with that excess EEF) ancestry, while o1 does not. Sintashta_MLBA_o1 is pretty much Yamnaya + WSHG.
The Siberian ancestry of Sintashta_MLBA_o1 is a bit shifted toward Botai compared to West Siberia N, while o3 does not. (Sintashta_MLBA_o3 accepts 0 Botai if West Siberia Neolithic is present in the same run, while o1 still takes a lot of Botai even then.)

PF said...

Can you make sense now where Anatolia_ChL and Armenia_ChL fit into the scheme of things?

Ha! This was exactly my first intention when I saw the post. Here's a quick run but I don't think it elucidates much.

[1] "distance%=3.1516"

Anatolia_ChL

Tepecik_Ciftlik_N,61.2
Darkveti-Meshoko,30.6
Progress_Eneolithic,8.2


Very much looking forward to the Dzudzuana and Anatolia_HG genomes to be released, but I also think we need Mesopotamian samples to make final sense of everything. It's really interesting what was there before/during the Neolithic; I really think it ended up having an important influence going in many directions. Should be something like CHG, but not quite...

To take it a step back a little, it seems that CHG is intrusive to the Caucasus ( = Dzudzuana-like + ANE-like), so we may see all sorts of CHG type populations north of and around the Caucasus, somewhere along the Steppe, maybe even south and east of the Caspian. Who knows. I mean, damn, there is something CHG-like in MA1!

Very general big picture for Paleolithic: I think the dominant dimension is influenced by ANE interactions across central Eurasia, with ANE + WHG forming EHG, ANE + Dzudzuana forming CHG, and ANE + Dzudzuana + extra Basal forming Iran_N. I digress a bit, though I do think it's relevant to the questions at hand.

A said...

@Samuel Andrews.

Thanks. Presumably their hunter gatherer ancestry is pulling them away somewhat from Yamnaya. I was wondering more about which population would be closest to Yamnaya overall. That seems to be 'Russians'.

Leron said...

So CHG and steppe gets brought up a lot, but where do the Sumerians and their predecessors fit into this? Are they indirectly represented by Iran_N?

Drago said...

@ Davidski

''Can you make sense now where Anatolia_ChL and Armenia_ChL fit into the scheme of things?''

We've basically done what we can with the data, even with more samples. The key now is draping the genetic data on the physical landscape & settlement, & understanding the exact sequence of population shifts. One cannot propose a population model with stats alone

Drago said...

@ Sam

''Steppe_Eneolithic & Caucasus_Eneolithic were separated by the Caucasus''

You mean the mountains ? That was not the barrier, because there is south Caucasian ancestry in Meshoko, Majkop, steppe-Majkop outlier, and Ukraine Yamnaya-outlier (maybe even some of the Vucedol), which are all north of the Caucasian mountains.

Davidski said...

@Dragos

The Caucasus was indeed a barrier, but obviously not exactly like an impenetrable wall.

Rather, West Asian populations coming from the south weren't adapted to the steppe very well, so they mostly stayed in the mountains.

Conversely, Eastern Europeans from the steppe weren't adapted to the mountains, so they mostly stayed on the steppe.

Only in some instances did this change, usually in times of drought.

Samuel Andrews said...

@Leron,
"So CHG and steppe gets brought up a lot, but where do the Sumerians and their predecessors fit into this? Are they indirectly represented by Iran_N?"

Sumerians were for sure mixed in someway not descended from a single Neolithic population. Samples from Iran contemporary to Sumerians are mostly Iran_N but also have Caucasian & Levant ancestry.

Drago said...

@ Davidski,
Right, so it was not the mountain range itself (as there were confluent zones with Caucasus ancestry - Meshoko, Majkop - in the northern foothills); but movement onto the steppe itself was limited to odd inviduals (for various reasons).
Similar but conversely, steppe movement to south of the Caucasus (as early as Chalc. Armenia 4300 BC); however was this a transient epidose ? The successding K-A Bronze Age seems a new movement from southern areas (lake Van, Urmia, NW Iran) back north and west.

A said...

When is someone going to try to get DNA from Sumerians? Kind of baffling that there's nothing so far.

Samuel Andrews said...

PG2004 is the best representative of pre-Yamnaya population. Still some eastern European hunter gatherer that PG2004 can't explain. CHG levels in northern Europe might only be 13-16% but "eastern hunter gatherer" is 30-35%. The importance of CHG/IranN ancestry is blurring more and more.

3.3267"

Yamnaya_Kalmykia

Progress_Eneolithic:PG2004,83.8
Globular_Amphora_Poland:I2403,7.5
Sweden_N,5.2
Ukraine_N,3.4
EHG,0.1


3.0157"

Sintashta_MLBA

Progress_Eneolithic:PG2004,55
Globular_Amphora_Poland:I2403,23.5
Sweden_N,9.9
Ukraine_N,6.1
Globular_Amphora_Ukraine:ILK002, ILK003,3.8
EHG,1.7

4.2725"

Icelandic

Progress_Eneolithic:PG2004,36.3
Globular_Amphora_Ukraine:ILK002, ILK003,34.9
Ukraine_N,10.1
Sweden_N,9.8
Globular_Amphora_Poland:I2403,8
Romania_HG,0.9

Davidski said...

@Dragos

Right, so it was not the mountain range itself (as there were confluent zones with Caucasus ancestry - Meshoko, Majkop - in the northern foothills); but movement onto the steppe itself was limited to odd individuals (for various reasons).

The height of the Caucasus and lack of safe passes across it would have been the initial problems, because they prevented a large number of people from the south from moving across it to the edge of the steppe.

It seems that without this critical mass of people there weren't enough resources and motivation to really have a go at taming the steppe, which was already home to a very aggressive population that probably didn't want to be colonized.

Cy Tolliver said...

What exactly is the relationship between CHG and Iran_N to begin with? Are they just two populations with mutual admixture between each other, or do they form a deeply rooted clade together but without much recent mixture between them? And how comparable are their levels of Basal Eurasian ancestry (or ANE for that matter)?

Drago said...

@ Matt

''Steppe_Maykop probably represent the least divergent of a set of western steppe populations, who were not quite purely paleolithic survivors but close to, which were overlian by expansions of Steppe_EBA toolkit and people, and absorbed on a low level.''

Chronology would stipulate that Steppe Majkop represent a migration of Botai/ W Sib. type groups into the north Caucasus. Their ''toolkit'' is that of Majkop.
It was they who absorbed earlier, Piedmont-steppe groupd.

a said...

Cy Tolliver said...
What exactly is the relationship between CHG and Iran_N to begin with? Are they just two populations with mutual admixture between each other, or do they form a deeply rooted clade together but without much recent mixture between them? And how comparable are their levels of Basal Eurasian ancestry (or ANE for that matter)?

26K YBP +/- Dzudzuana[Georgia] have not been released; however the authors of 'Paleolithic DNA from the Caucasus reveals core of West Eurasian ancestry" site the connection between Iran/CHG samples.


" We finally show that the Dzudzuana population contributed the majority of the ancestry of post-Ice Age people in the Near East, North Africa, and even parts of Europe, thereby becoming the largest single contributor of ancestry of all present-day West Eurasians."

https://www.biorxiv.org/content/10.1101/423079v

Andrzejewski said...

@Aram "he "North Caucasus" label in the paper is somewhat misleading, because it refers to different things.
North Caucasus as a geographic region and to the North Caucasian archaeologic culture. The other name for this culture is a Kubano Tersk culture. This is a EMB culture from ~3000-1500 BC period. It was known to have a more northern origin and now this is genetically confirmed. It is a simple offshoot of Yamna. With the same y dna.It is a steppic culture and has no presence in mountains."

Very good! You are acknowledging a North -> South axis of migration here from Steppe into Caucasus.

Andrzejewski said...

@Matt "Robustly then the later populations certainly have at least both Levantine and Anatolian components coming in. Looks like particular a Levantine/Mesopotamian flow as related to Levant_Chl? (Since later Levant+Arabian Peninsula pops seem most close relative to prediction/correlation of CHG+Barcin). Relating to Davidski's ideas about Uruk expansion impacting NE?"

Hold on a sec, Matt. Are you saying that modern Bedouin/Arabian/Levant populations are something between Barcin (Antalia_N) and CHG? I thought that they were mainly Natufian/Levant_N?

Wasn't the contribution of EEF to the formation of Iron Age Eastern Med populations more negligible compared to that or either CHG or especially Iran_Chl?

Going off on a tangent here, would you and @Ric Hern attribute the prevalence of blue eyes in Neolithic Levant population to a huge infusion of Anatolia_N (Hurrians?) into the Levant (Peki'in cave) as per the study done by Iosif Lazaridis?

Andrzejewski said...

@Samuel Andrews "Georgian, Abhkasian, probably several other modern Caucasians are continuation of Caucasus Chalcolithic (maybe Neolithic).

IMO, there's for sure direct Caucasus ancestry in the whole Middle East. But may not be "recent" because it already shows up in Iran & Anatolia in 4500 BC.

Kura Axes, Iran Chalcolithic, Armenia EBA are very mixed populations. A population mixed between Caucasus, Iran_N, Anatolia_N, Levant_N could be the secret successful Chalcolithic Y DNA J group people look for that spread across Middle East & into Balkans."

Okay:

1. Would you agree with me that modern Caucasus populations are mostly Anatolia farmer based than original CHG, and that perhaps their languages are more closely related to the Aegean Tyrresian proposed language macro-family than to the Kotias Klde 1 or Satsublia cave dwellers?

2. Are modern Caucasus dwellers different genetically from modern Eastern Med/West Asia inhabitants (Arabic speakers etc.), being both a melange of Anatalia_N, CHG, Iran_Chl and Levant_N (Natufians)? Or are the ratios of admixture components different in the different regions?

K33 said...

[i]a said: What exactly is the relationship between CHG and Iran_N to begin with? Are they just two populations with mutual admixture between each other, or do they form a deeply rooted clade together but without much recent mixture between them? And how comparable are their levels of Basal Eurasian ancestry (or ANE for that matter)?[/i]

My understanding is that Iran_N and CHG are both stabilized populations resulting from late Paleolithic admixture between:
Villabruna/Euro_HG + Basal Eurasian + ANE + ENA

However, the proportions of these components are different. The chief difference being that Iran_N has more Basal and less Villabruna/Euro HG ancestry relative to CHG.

Andrzejewski said...

@Samuel "Steppe_Eneolithic & Caucasus_Eneolithic were separated by the Caucasus. Steppe_Eneolithic led to rise of Yamnaya-like groups who became fundamental to European ancestry.

The Caucasus separation, also really important shows there wasn't enough exchange between north & south for IE languages to have originated south of the Caucasus. Yamnaya really was a home-grown European population native to land North of the Caucasus & West of Ural mountains."

Thank you for acknowledging that! I am thinking that after LGM, CHG forager groups ventured to the Southern PC Steppes (10,000 BC onward?) and then following 5000BC and later EHG groups started venturing south/west, met these groups and mixed with them via the female mediated bridal exchange. CHG in the Steppe probably by then spoke unrelated languages to ones spoken in the Caucasus.

Andrzejewski said...

Slumberry "Based on my eralier nMonte runs there are two mayor differences between o1 and o3.
Sintashta_MLBA_o3 accepts some actual Sintashta (and with that excess EEF) ancestry, while o1 does not. Sintashta_MLBA_o1 is pretty much Yamnaya + WSHG.
The Siberian ancestry of Sintashta_MLBA_o1 is a bit shifted toward Botai compared to West Siberia N, while o3 does not. (Sintashta_MLBA_o3 accepts 0 Botai if West Siberia Neolithic is present in the same run, while o1 still takes a lot of Botai even then.)"

Does it indicate an admixture with Botai? If so then it means that Andronovo has assimilated rather than displaced Botai in Kazakhstan.

Andrzejewski said...

Phillipe "Thanks. Presumably their hunter gatherer ancestry is pulling them away somewhat from Yamnaya. I was wondering more about which population would be closest to Yamnaya overall. That seems to be 'Russians'."

Not necessarily. Norwegians are 50% Yamnaya.

K33 said...

Honestly though, if Maykop is not responsible for the rise in CHG from Khvalnysk to Yamnaya, what's the current next best guess?

What about female-mediated admixture from the Kemi-Oba Culture? The time and the place would seem to work.

Crimea seems as good a place as any to wind up being a sink for CHG ancestry...

Andrzejewski said...

@Matt "Robustly then the later populations certainly have at least both Levantine and Anatolian components coming in. Looks like particular a Levantine/Mesopotamian flow as related to Levant_Chl? Relating to Davidski's ideas about Uruk expansion impacting NE?"

I don't remember Davidski ever mentioning the Uruk expansion impacting anything.

If at all, it was me who brought up a study about the prehistory of the Caucasus last week, with the stated goal of refuting any large scale population replacement in the Caucasus, especially a colonization of Uruk (Sumerian) dwellers.

My point was that neither one of the large macro-families endemic to the Caucasus has anything to do with the Sumerian language, who is widely considered to be isolate but which some linguists link to Sino-Tibetan.

A commentator named Ygor contradicted me, cautiously saying that the fact that none of the CURRENT modern languages in the Caucasus has nothing to do with SUMERIAN still doesn't rule out an Uruk expansion based on the possibilities that the language either died out, that the migrants were absorbed into autochthonous populations or that even the migrants weren't Sumerians but Ubaidians.

It IS known that the Sumerian language has some strong substrate tentatively termed "Ubaidian" or "Euphratean" but no one knows anything about their linguistic or generic affiliation. Some even thought they spoke some para-PIE language, go figure!

But, now that we know that Maykop did NOT contribute major genes nor language to the Steppe, it is largely irrelevant to our topic of conversation by any means.

However, it is known that after 6500YBP Anatolia_N populations moved to the Caucasus; maybe Kartvelian is an EEF language rather than CHG one?

And the pendulum swung the other way too, with a CHG-rich EEF population moving all over Europe, the Middle East and Iran and colonizing huge swaths.

Andrzejewski said...

@Lenny Dykstra "Honestly though, if Maykop is not responsible for the rise in CHG from Khvalynsk to Yamnaya, what's the current next best guess?"

I've just posted it! Post-LGM expansion into the Steppe after 10,000BC. They were native to that part of the Steppe.

Folker said...

@Lenny
The real difference between Khvalynsk and Yamnaya is EEF ancestry, not a rise of CHG related ancestry

Davidski said...

@Lenny Dykstra

Honestly though, if Maykop is not responsible for the rise in CHG from Khvalnysk to Yamnaya, what's the current next best guess?

We don't need to guess. People with way more CHG than Yamnaya were already on the steppe before even Maykop got going, and we have their genomes.

Progress_Eneolithic PG2001
Progress_Eneolithic PG2004
Vonyuchka_Eneolithic VJ1001

They were native to the steppe and probably the descendants of foragers living just north of the Caucasus.

K33 said...

"We don't need to guess. People with way more CHG than Yamnaya were already on the steppe before even Maykop got going, and we have their genomes."

Oof, I didn't notice that in the paper. They said the Progress people were "very similar to Khvalynsk" but looking at their own admixture bar graph they clearly do have much higher CHG than Khvalnysk.

(Using deliberately "distal" fixed sources to keep it simple):

--------------------------
[1] "distance%=3.4486"
Khvalynsk_Eneolithic
EHG,84.2
CHG,14.8
Boncuklu_N,1

[1] "distance%=4.3806"
Progress_Eneolithic
EHG,53.4
CHG,35.6
Ganj_Dareh_N,11

[1] "distance%=5.5814"
Yamnaya_Samara
EHG,58.6
CHG,33.8
Boncuklu_N,6.6
Ganj_Dareh_N,1

[1] "distance%=5.9063"
Yamnaya_Ukraine
EHG,56.4
CHG,35.6
Boncuklu_N,7.2
Ganj_Dareh_N,0.8

[1] "distance%=5.3991"
Yamnaya_Caucasus
EHG,54
CHG,36
Boncuklu_N,7.4
Ganj_Dareh_N,2.6
------------------------------

Kind of shitty distances with this approach, but the Yamnaya samples seem closer to Progress (Eneolithic Piedmont Steppe) than to Khvalynsk tbh. It's just b/c they, like Progress, are much more CHG-shifted than Khvalynsk.

Let's try more proximate models. FWIW, Steppe Maykop was rejected here as a source in all these:

-------------------------------
[1] "distance%=3.1437"
Yamnaya_Samara
Progress_Eneolithic,66.2
Khvalynsk_Eneolithic,25.4
Globular_Amphora_Ukraine,8.4

[1] "distance%=4.0583"
Yamnaya_Ukraine
Progress_Eneolithic,69.8
Khvalynsk_Eneolithic,20.8
Globular_Amphora_Ukraine,9.4

[1] "distance%=2.8551"
Yamnaya_Caucasus
Progress_Eneolithic,80.6
Khvalynsk_Eneolithic,9.8
Globular_Amphora_Ukraine,9.6
--------------------------------

Is there any core of truth to these runs??? That the Piedmont Steppe could have played an outsized role in the formation of the Yamnaya genotype? Archaeologically, I'm aware that Khvalnysk is supposed to have been sort of the granddaddy of PIE/Yamnaya etc...

K33 said...

Vonyuchka_Eneolithic is actually slightly more CHG-shifted than Progress (53% CHG/Iran_N-related vs 47% for Progress), but the results are similar. None of the Yamnaya pops accept more than 25-30% ancestry from Khvalynsk when Vonyuchka or Progress are added to the picture (along with GAC)

Gökhan said...

Well is the data from paleo caucaus from this paper released?
https://www.biorxiv.org/content/10.1101/423079v1

If there were a native CHG like pop in eneolethic steppe it might have paleothic backround i suppose.

George Okromchedlishvili said...

Andrzejewski

"However, it is known that after 6500YBP Anatolia_N populations moved to the Caucasus; maybe Kartvelian is an EEF language rather than CHG one?"

Both Kartvelian and NWC are poor match for CHG language families but look good as farmer ones

Aram said...

Steppe itself was the main problem for farmers. Notice for Central Asians farmers there was no barrier yet they didn't move to Kazakhstan South Ural.
The same can be said for EEF farmers. Cucuteni Trypllia was present in Moldavia quite early but they were moving quite slowly to east.

Matt said...

@Lenny, yes, these samples look like the main ancestry of Yamnaya, because they are CHG rich. We don't know whether they represent a mix which was old, or fairly young, at their time.

Using a "pure" model of the two "southernmost" Piedmont_Eneolithic samples PG2001 and VJ1001 (as PG2004 has more EHG than the other two and Khvalynsk likely has Piedmont_Eneolithic steppe ancestry as a heterogenous mixing point for southern steppes peoples and Samara locals): you should get a model of about EHG:SouthSteppe:MN_Europe - 23:67:10.

@dragos: "Chronology would stipulate that Steppe Majkop represent a migration of Botai/ W Sib. type groups into the north Caucasus. Their ''toolkit'' is that of Majkop."

This implies you seem to have a great deal of knowledge about the shared technology and subsistence practices between the Steppe_Maykop and Mountain Maykop groups and the genesis of the former. I could not find much in English about this, as really all that seems to have made it over to English is Shishlina's subsistence isotope ratio studies, so could not comment either on that much.

Since you have all this information to hand, anything you could enlighten me on about these mysterious people? Please point me to English language references if possible.

The dominant ancestry for Steppe_Maykop in models is Piedmont_Eneolithic of course, although it's probably more like 50:50 West_Siberia:Piedmont Eneolithic, if you use PG2001 and VJ1001, and disregard AY2001 who may have genuine ancestry from the mountain region.

This may represent a "recent" mix of Piedmont_Eneolithic and West Siberia, or, just as likely, a long running cline through the North Caspian deserts, if we following Davidski's theory that Piedmont_Eneolithic represents an old presence of fairly high CHG in south steppe region (esp. following suggestions Kelteminar are probably autosomally similar to West_Siberia_N).

Archaeologically, I don't think there's certainly for Botai any suggestion that they expanded anywhere. Seem to be only a local, sedentary culture of horse farmers, farming horses for meat and milk and living in sedentary village(s). Until perhaps they get pushed east by later expansions...

@ andrej: Re; Uruk/Ubaid and Davidski's theory on spread of J through ME - http://eurogenes.blogspot.com/2017/05/a-homeland-but-not-homeland.html

Samuel Andrews said...

Chalcolithic Iran had both Levant & Caucasus admixture.

1.79"

Hajji_Firuz_ChL

Darkveti-Meshoko,32.3
Levant_ChL,28.2
Ganj_Dareh_N,26.1
Barcin_N,13.4
Progress_Eneolithic:PG2001,0
Progress_Eneolithic:PG2004,0
Vonyuchka_Eneolithic,0
CHG,0
Levant_N,0
Tepecik_Ciftlik_N,0
Botai,0
AfontovaGora3,0

Kura_Axes had admixture from Iran.

2.3695"

Kura-Araxes_Kaps

Darkveti-Meshoko,49.6
Seh_Gabi_ChL,34.8
Barcin_N,11.2
Progress_Eneolithic:PG2001,3.5
Vonyuchka_Eneolithic,0.9

Samuel Andrews said...

@Phillipe,
"@Samuel Andrews.

Thanks. Presumably their hunter gatherer ancestry is pulling them away somewhat from Yamnaya. I was wondering more about which population would be closest to Yamnaya overall. That seems to be 'Russians'."

Uralic ancestry creates psuedo Yamnaya affinity. Many Russians have significant Uralic Karelian-like ancestry. In PCAs this pushes them towards Yamnaya.

Karelians have a lot of Saami ancestry. Saami have a lot of Artic Siberian ancestry. In this Artic ancestry is a lot of ANE. It's this which pulls Saami, therefore Karelians, therefore some Russians east towards Yamnaya. In total Yamnaya ancestry percentage, Russians have less than other eastern Europeans because of Uralic admixture.

Davidski said...

@All

Here are two more models for the Sredny Stog R1a-M417 sample. In the first model, Darkveti-Meshoko is rejected. When I remove this sample, the model looks awesome.

Sredny_Stog I6561 doesn't have any ancestry from the Caucasus.

Sredny_Stog_I6561
Darkveti-Meshoko -0.003±0.070
Progress_Eneolithic_PG2001 0.405±0.070
Ukraine_N 0.161±0.035
Varna 0.437±0.047
chisq 7.650
tail prob 0.662933
Full output (infeasible)

Sredny_Stog_I6561
Progress_Eneolithic_PG2001 0.419±0.041
Ukraine_N 0.163±0.027
Varna 0.418±0.032
chisq 6.866
tail prob 0.809861
Full output

Drago said...

@ Matt

I would presume we agree that Eneolithic Piedmont-steppe samples, like Progress, are older than Steppe Majkop. The former date to c. 4300 cal BC; and the latter to after 4000 BC. According to the chronological summary in Fig 2b, there might even be a settlement hiatus, which might be due to technical reasons or a true disruption broadly linking in with that seen in Varna, Suvorovo, etc. Whatever the case, Steppe Majkop follow the Progress/ VJ individuals. As such it would be reasonable to see what modelling Steppe Majkop does with preceding Progress.:

Steppe_Maykop
West_Siberia_N 43.5%
Piedmont_Steppe 39.2%
Sarazm_Eneolithic 9.5%
Samara_Eneolithic:I0122 6.2%
Levant_N 1.6%
Varna 0%
Darkveti-Meshoko:I2056 0%
d 1.6%

So as you said, essentially a 50/50% mix.

So the quesiton is to your suggestion ''This may represent a "recent" mix of Piedmont_Eneolithic and West Siberia, or, just as likely, a long running cline through the North Caspian deserts, if we following Davidski's theory that Piedmont_Eneolithic represents an old presence of fairly high CHG in south steppe region (esp. following suggestions Kelteminar are probably autosomally similar to West_Siberia_N).

However, all older samples - whether Meshoko or Progress don't have much in the way of significant Siberian -like ancestry.

Eg
Darkveti-Meshoko:I2056
CHG 63.7%
Tepecik_Ciftlik_N 32%
Ganj_Dareh_N 4.3%
EHG 0%
d 2.9%

Piedmont steppe (avg)
EHG 42.3%
CHG 38.9%
Ganj_Dareh_N 14.9%
West_Siberia_N 3.9%
d=4.5%

At least according to the present data, the conclusion fabours that - whatever its origin- there is a sharp & significant rise in Siberian ancestry after 4000 BC in the north Caucasus steppe.

Whilst this could indeed represent a relictual Paleolithic origin, this doesn't seem parsiminous for several reasons (1- the Imeratian origins of the north Caucasian epipaleolithic; 2- the Chalcolithic tool-kit of steppe majkop; 3- the ecologically volatile and anthropologically poorly documented record of the Khvalynian desert/ north Caspian steppe). Rather, it seems like a distinct and deliberate affair, associated with some conforming Y-Hgs (eg Q1, IIRC).


Previous thoughts thought Steppe Majkop was a colonization process of the Mountain Majkop folk moving north. So I guess, much of the said here is probably news to everybody, in all modesty. Finer details & synthesis will have to wait.

Matt said...

@Dragos, I'm not arguing that Steppe_Maykop represents paleolithic continuity in exactly same region (of course not), but poss. not migration of some pure West Siberian / Botai like folk plus purely Piedmont_Eneolithic folk, rather than some more proximate population of semi-shrub desert / desert steppe zone culturally integrating into some exchange with Maykop. (Genesis of this phenomenon I wouldn't claim to understand, and probably would require tight reading of archaeology around the coast of Caspian Sea and associated floodplains, possibly involving sites that are currently underwater).

That is Steppe Majkop does not "represent a migration of Botai/ W Sib. type groups into the north Caucasus" but rather larger migration of some population already intermediate "W Sib. type groups" and probably not Botai, and Piedmont_Eneolithic.

Re; non-conforming y-dna, note R1b variants in Botai, Q variants in (low n) Khvalynsk samples. Not wise to over-generalise migration from Eastern Kazkhastan from this.

Matt said...

Comparing variability with the Piedmont_Eneolithic samples.

1. Population Averages - top 15 most residually closer to PG2004 over PG2001 (in order): Sintashta_MLBA_o3, EHG, Srubnaya_MLBA_o, Khvalynsk_Eneolithic, West_Siberia_N, AfontovaGora3, Oy_Dzhaylau_MLBA_o1, Sintashta_MLBA_o1, Ukraine_Mesolithic, Latvia_MN, Botai, Dali_EBA, Steppe_Maykop, MA1, Comb_Ceramic

2. Population Averages - top 15 most residually closer to PG2001 over PG2004 (in order): Darkveti-Meshoko, North_Caucasus_MBA, Kura-Araxes_Kaps, Kura-Araxes_Velikent, Maykop_Novosvobodnaya, Georgian_Imer, Maykop_Late, Sumbar_LBA, Seh_Gabi_ChL, Kura-Araxes_Talin, CHG, Kura-Araxes_Kalavan, Hajji_Firuz_ChL, Georgian_Laz, Abkhasian

3. Population Averages - top 15 most residually closer to PG2001 over VJ1001* (in order): Boncuklu_N, Mentese_N, Koros_N, Greece_N, LBKT_MN, Beaker_Rhine-Main, Varna, LBK_N, Peloponnese_N, Tiszapolgar_ECA, TDLN, Balkans_N, Beaker_France_no_steppe, Starcevo_N, Vinca_MN

*disregarding an odd signal where Papuan pops closer to PG2001 over VJ1001 which I assume is marginal error.

4. Population Averages - top 15 most residually closer to VJ1001 over PG2001 (in order):: Vonyuchka_Eneolithic, Ganj_Dareh_N, Abdul_Hosein_N, Sarazm_Eneolithic, Hotu_HG, Wezmeh_Cave_N, Seh_Gabi_LN, Tepe_Anau_Eneolithic, Geoksiur_Eneolithic, Parkhai_Eneolithic, Parkhai_EBA, Parkhai_MBA, Parkhai_LBA, Gonur1_BA_o, Parkhai_LBA_o

5. Population Averages - top 15 most residually closer to PG2004 over VJ1001 (in order): Khvalynsk_Eneolithic, Sintashta_MLBA_o3, EHG, Srubnaya_MLBA_o, AfontovaGora3, West_Siberia_N, Oy_Dzhaylau_MLBA_o1, Maitan_MLBA_Alakul_o, Sintashta_MLBA_o1, MA1, Ukraine_Mesolithic, Botai, Latvia_MN, Comb_Ceramic, Dali_EBA

6. Population Averages - top 15 most residually closer to VJ1001 over PG2004 (in order): Vonyuchka_Eneolithic, Tepe_Hissar_ChL, Parkhai_LBA, Sumbar_LBA, CHG, Seh_Gabi_LN, Parkhai_MBA, Wezmeh_Cave_N, Tepe_Anau_Eneolithic, Parkhai_Eneolithic, Ganj_Dareh_N, Abdul_Hosein_N, Parkhai_EBA, Geoksiur_Eneolithic, Seh_Gabi_ChL

......

In the case of PG2004 vs PG2001+Vonyuchka, the difference is fairly solid, as established. Pair differences within PG2001+Vonyuchka could be within noise, certainly wouldn't say large enough to suggest a significant population difference and nothing suggesting cause to change ideas on anything.

There just aren't really enough samples here to try and do cline extension of within group variation with any confidence...

Distances of three samples: https://i.imgur.com/9MCEAt5.png

All these samples and steppe_maykop seem closer to Yamnaya in G25 than EHG+CHG would predict. But I can't see any signal that Piedmont_Eneolithic are closer to anyone else that EHG+CHG would predict, in particular.

----

7. Population Averages - top 15 most residually closer to PG2001+VJ1001 over Khvalysnk pair: CHG, Parkhai_MBA, Parkhai_Eneolithic, Geoksiur_Eneolithic, Sarazm_Eneolithic, Tepe_Anau_Eneolithic, Sumbar_LBA, Parkhai_EBA, Tepe_Hissar_ChL, Parkhai_LBA, Parkhai_LBA_o, Vonyuchka_Eneolithic, Namazga_Eneolithic, Wezmeh_Cave_N, Gonur1_BA

8. Population Averages - top 15 most residually closer to Khvalynsk pair over PG2001+VJ1001: Khvalynsk_Eneolithic, EHG, Latvia_MN, Ukraine_Mesolithic, Sintashta_MLBA_o3, Ukraine_N, Comb_Ceramic, SHG, Narva_Estonia, Latvia_HG, Romania_HG, Narva_Lithuania, Baltic_HG, Iron_Gates_HG, West_Siberia_N

Matt said...

Prompted by some of the above, having a closer look at the Maykop samples in the paper, in roughly chronological order (breaking up into separate posts):

- Maykop: I1720 (5300 BP) and OSS001 (5570 BP). Both from southeast of the Maykop range on the north face of mountain flank and in PCA breaking from the main cluster in trailing off towards Anatolia (one more than the other; hard, not sure which).

Neither in G25.

- Maykop Novosvobodnaya: I6268 (5564 BP), I6267 (5434 BP), I6272 (5200 BP), I6266 (5200 BP). Site of Klady, northwest of the Maykop range, as far as that goes on the mountain flank itself (but not as ultimately far NW as that goes, into the Kuban Steppe). Pretty solidly in the Cacasus cluster, and don't appear particularly enriched for CHG in analysis (unlike Darkveti-Meshoko or North Caucasus MBA)... but note in PCA, 2 samples are trailing away to Iran_Chl.

2 of the samples in G25: I6266, I6272. These are both the later two of the samples. Neither looks obviously more like they are outlying in the ADMIXTURE by Wang. But it would be interesting if either looked more Seh_Gabi like in tests on G25.

"In archaeology the Klady cemetery is regarded as the most vivid example of the relationship between the European, Caucasian and Ancient East cultures in the Early Bronze Age."

Matt said...

(cont).

- Steppe Maykop:

1. Marinskaya site: MK5005 (5455 BP): The earliest and geographically the southmost of the Steppe_Maykop samples. Looks out of all Steppe_Maykop least distinct from Piedmont_Eneolithic in ADMIXTURE (little red East Asian component).

There are only 4 Steppe Maykop samples in Wang's PCA (excluding the 2x outliers), and judging by the correlation of positions between G25 and Wang's PCA, MK5005 is none of these and she's not in either.

Her SNPs are not very good at 85910, which is a shame as she is in a sense one of the quite interesting ones among the Steppe Maykop, both because of her early time, low apparent Siberian affinity, being from the same site as Late Maykop samples, and the one positioned most closely to Progress and Vonyuchka Eneolithic. If there were any Steppe_Maykop who were straightforwardly Piedmont_Eneo people absorbed into the Maykop culture, it would be her. However she still does have some red East Asian component.

2. Sharakhalsun site: SA6001 (5444 BP), SA6013 (5180 BP), SA6004 (5170.5 BP): Samples from the northmost and eastmost of the Maykop range, outside the mountain flank, closest to the desert and North Caspian. All three are in G25.

SA6001 and SA6004 are the Steppe Maykop with the highest levels of red East Asian and Aqua components.

SA6013 is one of the two samples classified as an outlier by Wang. This sample is between Steppe Maykop and Caucasus in Wang's PCA and the less Anatolian of the two. The other Steppe_Maykop Outlier is IV3002, with more plentiful Anatolian component, and is between Caucasus and Yamnaya on the PCA.

3. IV3002 (as above) is from further west and dated 5206 BP, and is furthest west of the Steppe Maykop at Ipatovo. IV3002 is not in G25.

4. Aygurskiy site: AY2003 (5455 BP), AY2001 (5271). Intermediate site to Ipatovo (most west) and Sharakhalsun (most east) sites. AY2003, the earlier sample sits with the high Siberian samples from Sharakhalsun, while AY2001 the later sample is the 'intermediate' sample who ends up closest in Wang's PCA and G25 to the Piedmont_Eneolithic (though being clearly different). Both in G25.

Matt said...

(cont.)

- Late Maykop: MK5008 (5185.5 BP), MK5004 (5171 BP), MK5001 (5141.5), I11132, I11131, I11133. The latter three samples seem to have changed labelling between ADMIXTURE and supplement, where they are described as SIJ001 (5125.5 BP), SIJ002 (5173.5 BP), SIJ003 (5174 BP. MK samples are from Marinskaya while SIJ samples from Sinyukha, slight NW of Klady but still on mountain flank. A couple of the samples may have more Anatolian than others, but otherwise these look like very "normal" Caucasus cluster samples. MK5004 and SIJ003 are both in G25 - others aren't, but seems like no great loss as doesn't look like anything too special going on with them based on Wang's PCA and ADMIXTURE.

....

Overall, does seem like there is a general tendency where earliest Maykop samples include flow in from external sources and more homogenised later on, and that within the Steppe_Maykop, there is a (albeit not monotonic) tendency to decreasing Siberian related ancestry at sites away from the east and with increasing time.

Frustratingly, there aren't actually any samples representing Maykop intrusion into the west end of the steppes north of the Caucasus range, near the Black Sea and lower Don, where it is most likely that if they were interacting with Sredny-Stog at all, it was here. The Steppe Maykop sampled seem largely at the northeastern limits of the Maykop range, which shade into a quite different, arid desert ecozone that persists around the North Caspian.

It's not clear to me that Steppe_Maykop is well thought of as a unitary phenomenon, rather than separate Maykop interactions with populations at the limits of their range, which would be different in the east and nw of that range..

Andrzejewski said...

@George O. "Both Kartvelian and NWC are poor match for CHG language families but look good as farmer ones"

I'm glad you agree with me on that one, but what's the evidence? You know that NWC is very different than Kartvelian.

Would you characterize NE Caucasus and/or Hurro-Urartian as a farmer language or a CHG one?

Matt said...

With above series of posts, just to say, I'm not saying that this paper doesn't indirectly tell us a *lot* about the likely lack of much demographically significant marriage interaction between Maykop and the plausible ancestor cultures of Yamnaya.

Indirectly we can tell a lot through the Piedmont Steppe samples and the Maykop samples - that Maykop+EHG is not really a parsimonious model for Yamnaya compared to Piedmont Steppe+Ukrainian Neolithic cultures.

Just that there don't seem like there are really any Maykop sites in the plausible contact zones around the Black Sea and nearer to the Don (and there should have really been some judging by most of the maps floating around on the web and in the paper) that can really tell us as much directly as we'd prefer.

Andrzejewski said...

@Matt "Archaeologically, I don't think there's certainly for Botai any suggestion that they expanded anywhere. Seem to be only a local, sedentary culture of horse farmers, farming horses for meat and milk and living in sedentary village(s). Until perhaps they get pushed east by later expansions..."

Here's what's fascinating about Botai and other ANE groups: the theory goes that they had tamed the horse way before our ancestors did it, but later pushed away by Andronovo. Now it turns out that Sintashta actually assimilated lots of Botai rather than displacing them.

But that's not the end of the story: According to the "classical" theory by Damgaard 2018 Botai got pushed to Mongolia. So, are they the ancestors of the ANE-rich, Yeniseian-like Okunevo groups who mixed with Afanasievo to create Tagar Culture, supposedly a Scythian material culture speaking an IE language? Scythian were for the most part, like Tocharians, very Europoid-looking.

Now, both Botai and Okunevo were found to be "genetically similar to American Indians". Another perplexing fact is that it turns out that the ancestors of the Huns, the Xiongnu were all found to belong to Haplogroups Q1a and Q1b with no exception, something like a dozen samples. Moreover, Chinese sources related a phrase in a Yeniseian language by a tribe belonging to the Xiongnu Confederacy, Jie/Kiat. The Jie were later exterminated by the Tang Empire.

What's the relevance to our discussion? Many:

1. Are the remaining Botai in Eastern Mongolia the ancestors of the Okunevo who later mixed with Indo-Europeans Afanasievo to create Europoid populations like the Scythian Tagar Culture?

2. Were the Huns/Xiongnu who invaded Europe Yeniseian-speakers rather than Turkic? That would explain the prevalence of Hap Q in many Slovaks, Hungarians, Romanians and Scandinavians, 2.5% - 5%: Huns, Avars, some Magyars (it's known that 25%-30% of the Hungarian language vocabulary is neither Uralic nor Turkic nor Indo-European).

3. Were the Botai and the largely similar BMAC (WSHG + some Iran_N ancestry) assimilated into the invading Andronovo Culture, and were some Zoroastrian religious and even vocabulary in Proto-Indo-Iranian either BMAC or Botai in origin?

4. Up until the 5th century when all of the sudden "Altaic" languages: Manchu, Turkic, Mongolian, Ruan Ruan make a sudden ascendence to the stage, most of Central Asia were apparently inhabited by "Paleo-Siberians" speaking ANE/Yeniseian-related tongues and some of them (in addition to the Eastern Hunter Gatherers in Steppe populations) may have contributed to the creation of Indo-European ethnogenesis.

5. The Tang Emperor ordered the extermination of the "Europoid looking" Jie/Kiat, who were also ANE and Yeniseian speaking. I wonder how many more "Europoid" looking ANE - Botai, Huns, Okunevo were roaming in prehistoric North East Asia, Central Asia and other parts of the continent, and how much of it was attributed to mixing with Yamnaya/Kurgan/Steppe derived populations (who were supposedly Cro-Magnon or Dolichocephalic Europoids, perhaps a minority were "Ladogan" or "Turanian" looking), and how many ANE populations were Caucasian-looking even without mixing with Indo-Europeans. In other words, how much of the ANE/Afontova Gora 3 contributed to the creation of the genetics vis-a-vis "looks"/appearance of the Proto-Indo-Europeans.

Slumbery said...

@Andrzejewski

"Does it indicate an admixture with Botai? If so then it means that Andronovo has assimilated rather than displaced Botai in Kazakhstan.

Note that Sintashta_MLBA_o1 is an outlier that have pretty much zero actual Sintashta ancestry. The Sintashta/Andronovo expansion trough the region was a mixture of displacement and mixture. What got displaced or assimilated however is not the outright Botai&related groups, because those did not live to see Sintashta. Before that Yamnaya groups had fused with Botai(like) groups. This gene flow went into both directions in some extent, Potakovka had significant WSHG ancestry above what Yamnaya and Poltakva had. This Yamnaya+Botai(like) population then slowly melted into Andronovo, but for example Srubnaya_MLBA_o is still totally Yamnanya + WSHG with zero excess western ancestry, so that process took time.

Anyhow, Botai culturally/archaeologically disappeared well before Andronovo formed, so technically we can say that it was not assimilated or displaced by Andronovo. It was assimilated by Yamnanya, then the resulting mixed population got displaced&assimilated by Sintashta/Andronovo, leaving a only a diminishing level of WSHG ancestry in the later steppe populations (Sarmatians had almost none of it, not even the Asian ones and even that low level was not necessarily from Botai).

Andrzejewski said...

@Slubery according to Damgaard 2018 there was almost zero gene flow between Botai and Yamnaya. He also claimed that Yamnaya-groups displaced Botai and their Paszerwski horses all the way to the Altai mountain.

Andrzejewski said...

@Phillipe @Samuel Andrews "Uralic ancestry creates psuedo Yamnaya affinity. Many Russians have significant Uralic Karelian-like ancestry. In PCAs this pushes them towards Yamnaya.

Karelians have a lot of Saami ancestry. Saami have a lot of Artic Siberian ancestry. In this Artic ancestry is a lot of ANE. It's this which pulls Saami, therefore Karelians, therefore some Russians east towards Yamnaya. In total Yamnaya ancestry percentage, Russians have less than other eastern Europeans because of Uralic admixture."

Uralics =/= ANE/EHG. Uralics have a different origin than Ancient North Eurasian and are therefore NOT even similar to EHG. Note N1c1 v. R1a/R1b/Q1a. Uralic languages and genetics are very different than EHG. I don't know why people keep on refering to Uralics as EHG-like. Even the Indo-Uralic theory has been debunked and discredited, based on the fact that there are so few common cognates like "water" or "name", and even them were proved to be loan words from PIE.

Andrzejewski said...

@Matt I need more material about the Mesopotamia -> Kura-Araxes -> Maykop.

Would you then claim that any modern Caucasus language family is a scion or offshoot of either Sumerian or "Ubaidian"/"Euphratean" instead of a CHG/KK1/Satsurblia language or even rather an Anatolia_N language?

I believe that the Uruk expansion if it even occurred and took place contributed very few to extant Caucasus populations, ancient or modern then did Iran_Chl or especially Anatolia_N. But I would rather see you post it here like the one I did (which I rebutted and refuted).

While we're on it, how would you explain the emergence of Semitic speaking people ~4300YBP? Note that it happened more or less at the same time of the resurgence of Iran_Chl populations in the Levant.

Slumbery said...

@Andrzejewski

"according to Damgaard 2018 there was almost zero gene flow between Botai and Yamnaya."

Not relevant for my claims, because I only used Yamnaya and Botai as ancestries, never claimed that there was significant gene flow between the representative Yamnanya and Botai groups in their own time.

He also claimed that Yamnaya-groups displaced Botai and their Paszerwski horses all the way to the Altai mountain.

nMonte runs clearly show the existence of a Yamnaya descendant + WSHG population well into Andronovo times, this is a fact. This population had a majority Yamnaya ancestry, so it is possible that Botai as a culture, language, etc. completely went down, together with their horses. It is also possible that this WSHG ancestry is not actually from Botai strictly speaking. But the Yamanaya+WSHG population is definitely there in a consistent presence of Sintashta-Andronovo-Srubnaya outliers. The cultural side of this is anybody's guess.

Matt said...

@andre, Nope, I don't see any of that Uruk/Ubaid genetic expansion, if it happened, associated with language shift and neither I believe would Davidski when he proposed it. Your other comments seem dumb and generally you usually seem to have a sketchy understand of the autosomal shifts, so don't really see much point spending time on them.

Andrzejewski said...

@All there was also an archeological culture called Volsovo which has contributed a lot in terms of genetics to the formation of Poltava, Strubna and Andronovo. It was regarded by some sources as "Proto-Uralic" whereas on the other hand others have claimed that it was a non-IE non-Uralic in its origin. Much of the non-Uralic substrate in Saami languages as well as some non-IE substratum in Baltic languages seem to have originated in Volsovo, and overall it seems to have ties to Kunda and Narva Cultures.

I therefore believe that the non-IE substratum in Balto-Slavic which is even responsible for the sound shifts and hence the formation of the Proto-B.S. branch may indeed be the non-IE non-Uralic relics of the Volosovo culture.

Andrzejewski said...

@Matt "Nope, I don't see any of that Uruk/Ubaid genetic expansion, if it happened, associated with language shift and neither I believe would Davidski when he proposed it."

No. I posted a study in earlier entries in which the authors attribute current and modern populations living in the Caucasus to this Uruk expansion period migrants (large scale!) in search of metallurgy. It alludes to the current Caucasus language families deriving from this migration rather than being indigenous (CHG) or even more documented migration from Anatolia_N into the Caucasus. As I had posted, and George O. agreed with me tentatively, both Kartvelian and NWC seem to be "farmer" tongues more in common with Aegean/Pelasgian and Etruscan/Raetians than either CHG or Sumerian/Ubaidian. I did not agree with him regarding NorthWest Caucasian and thus asked him to post a link to any reputable study substantiating it.

All I am asking you to do is back up your words with scientific article(s), the way I did myself.

Matt said...

@andrej, what statement do you believe I should back up? Davidski suggested there could be an expansion of an Iran_Chl like genetic component from Northern Mesopotamia coinciding with Ubaid period, and connected this with no language expansion. I offhand suggested that the changes in the Caucasus post-eneolithic seen in G25 may be related to the same process, if Davidski was correct about it.

I have said nothing about languages - in general I think it is a parlous game to try and demonstrate that any language in the ME or less a limited attested one in ancient S Europe or ME is linked to some particular component- and I generally don't read any of your comments as a rule, and not following any of your discussions with any other poster.

@Davidski, I'm pretty sure you've said that the samples that haven't translated over into G25 are due to low SNPs, and not much overlap, but is there any possibility of getting Lola:NV3001?

He's a late sample and represents a bounce back of Steppe_Maykop...

The interesting thing for me though, is that he is intermediate between Steppe_Maykop and others in most of the PCA, but it looks about equally plausible that he is between the extreme most HG Steppe_Maykop and the Caucasus as between Steppe_Maykop and Yamnaya/Steppe_EBA.

This matters, because his appearance (and his culture) coincides with aridisation in the steppe, and I would guess of groups more specialised to deal with arid climates.

So using G25 we could probably tell whether he represents a "re-emergence" of the old AY2001+SA6013+IV3002 style admixtures between Caucasus and "Steppe Maykop" extremes, or a "de novo" admixture of a people from the North Caucasus with the Catacomb culture gene pool. That could tell us something about the dynamics in the region that Wang couldn't address.

Supplement suggests the SNP coverage is pretty good - 439236 SNPs with Human Origins.

Chad said...

Hmm.. Not sure how the authors missed this...

left pops:
Steppe_EMBA
Progress_Eneolithic
Steppe_Maykop
Ukraine_N
Maykop_Novosvobodnaya

right pops:
Chimp
EHG
Iron_Gates
Ust_Ishim
CHG
Levant_N
Ganj_Dareh_N
Anatolia_N
Brazil_LopaDoSanto_9600BP
West_Siberia_N

numsnps used: 561112

best coefficients: 0.546 0.108 0.178 0.167
std. errors: 0.091 0.059 0.015 0.027

fixed pat wt dof chisq tail prob
0000 0 6 6.796 0.340124 0.546 0.108 0.178 0.167

Andrzejewski said...

@Matt " and I generally don't read any of your comments as a rule, and not following any of your discussions with any other poster."

Very heartwarming comments. Have a nice weekend too.

One thing *I* do which some posters neglect/fail to do is almost always post supporting evidence corroborating what I say.

Drago said...

@ Matt

“Not wise to over-generalise migration from Eastern Kazkhastan from this.”

I did not claim this represent a migration from east Kazakhstan ; but rather “”Steppe Majkop represent a migration of Botai/ W Sib. -type groups into the north Caucasus. “
I’m not actually sure where they’re from

Matt said...

@Chad, nice to see that works. I think on their part though it's possibly more of a "If models satisfy with 3 pops - Eneolithic_Steppe, GAC, Eneolithic_Caucasus and last only relevant for North_Caucasus and Late_North_Caucasus* - can we justify before peer review why we'd use a 4 pop model?".

If they'd found samples that looked like they could only be modelled by a four-pop interaction, like admixtured between SS and Maykop (the way Steppe_Maykop in eastern region shows admixture between probably Caspian and Maykop), they'd have probably put those models out there.

I'd guess it probably did to them that Ukraine_N and Anatolian ancestry via Maykop and Steppe_Maykop could probably balance out to what GAC provides in their model? They just didn't have a reason to go with what might look like a more complex 4-pop model without some justification for doing so. If any samples show up that look like they can only be explained that way, early enough, in a culture that looks ancestral to Yamnaya, with enough time to form a homogenous admixture...

*2 pop models with GAC and Eneolithic_Steppe basically work at above p>=0.05 for Yamnaya_Caucasus, Yamnaya_Ukraine, Yamnaya_Samara, Potapovka, Poltavka and Catacomb.

@Dragos, the only samples we have of a similar autosomal makeup to WSHG/Botai so far are from that neck of the woods. It's pure conjecture that we would have samples of the same makeup anywhere much to the west of there.

Though I agree the Steppe_Maykop samples (the ones who are not obviously admixed with some Eneolithic and post- Caucasus culture) are from elsewhere to the east, we're on even thinner ground if we're claiming we know they came from some unadmixed WSN like population. When we have no knowledge of how far the CHG component was distributed and it is plausible this has presence quite anciently north of the Caucasus.

There are no clear grounds to claim that we know any of AY2003, SA6001 and SA6004 are recent admixtures of WSN and the VJ1001+PG2001 individuals.

Drago said...

Archaeologically, I don't think there's certainly for Botai any suggestion that they expanded anywhere.”

Maybe not Botai specifically; however SiberianN type ancestry is found from Turan to IAMC herders to central steppe & now Fore-Caucasus .
Probably not very immobile

Matt said...

@Dragos, not in unadmixed form, which is the point. Certainly not unadmixed from IranN in samples we have, possibly admixed with CHG as well, not so clear. Turan some territory of purely WSHG people until Neolithic Iran expansion there, rather than long standing cline between IranN and WSHG.

Drago said...

@ Matt
There are no clear grounds to claim that we know any of AY2003, SA6001 and SA6004 are recent admixtures

Is there any LD analysis ?

The Nth Caucasus was a population sink with little evidence for Mesolithic or even Neolithic period habitation; few caves at best
Populations shifted all the time; eg even within the “ sedentary farmers” in Anatolia (several such horizons)
The SS then Majkop periods were drawing populations in. This is rather obvious by now

Drago said...

''Both Kartvelian and NWC are poor match for CHG language families but look good as farmer ones''

Can someone clarify this ? CHG were farmers too (eventually). What in linguistics stipulates they're from ANF rather than CHG. Seems a tad speculative , no ?

Andrzejewski said...

@Dragos "''Both Kartvelian and NWC are poor match for CHG language families but look good as farmer ones''

Can someone clarify this ? CHG were farmers too (eventually). What in linguistics stipulates they're from ANF rather than CHG. Seems a tad speculative , no ?"

I asked George the same thing.

Well, NWC speakers have more Steppe ancestry, which makes them look more European-shifted. If you noticed, Chechens (NEC speakers) and Kartvelians look more West Asian. On the other hand, Kartvelians are 40% Antaolia_N and haplogroup G and only 27% Haplogroup J.

If you noticed, Semitic speakers have Hap J too.

It seems like Kartvelian may have connections and cognates in the Aegean language family, NWC may be a relic from some non-IE CHG-rich population from the Steppe and Chechen may be CHG, as was Hurrian-Urartu.

Andrzejewski said...

@All my end goal is to find out whether PIE or pre-PIE was at its core an ANE/EHG language, a CHG one, a hybrid CHG-EHG or perhaps some language isolate which has developed independently on the PC Steppes without ANY link or relation to any other language.

Chad said...

3-way Ukraine, Maykop, and Progress mix looks as good.

Matt said...

Dragos: The Nth Caucasus was a population sink with little evidence for Mesolithic or even Neolithic period habitation; few caves at best

Doesn't really matter to the question of whether the Steppe_Maykop are, at their time, a recent admixture of Piedmont_Eneolithic and West_Siberia_N.

Thousands of years since the mesolithic period here. Thousands of years since the early neolithic in West Asia.

Is there any LD analysis ?

Nope, only for the obviously admixed Steppe_Maykop outliers (which was like 700 years before their time or something like that - it's in the supplement).

Matt said...

@Chad, probably worth to message the authors on that one if it fits as well, esp. if can accommodate all same pops as their final 3 pop model with similar p value.

Drago said...

@ Matt
We are talking about Caucasus not Syria or Israel
The “Neolithic” in the steppe and Caucasus is but 5000 BC; not “thousands of years earlier” . It’s imprtant to be aware of basic frameworks

So the admixture might not have occurred exactly in 3800 Bc; but “recently” (geologically). Over few hundred years ; not long after the local Neolithic, perhaps en route. The rapid expansion of Yamnaya & BB, where they remained relatively unadmixed, seem to be exceptional cases

Drago said...

@ Andre
Yep agree. Linking Chechens etc with EEF seems dubious , if that is what they mean by “farmers”

Matt said...

@Dragos, 7000 BP (5000 BCE) is, indeed, approximately 2000 years before the Steppe_Maykop samples. If there were, for some reason, reason to be confident that the expansion of CHG ancestry north of the Caucasus had to wait until the dawn of the neolithic in the Caucasus. Which there is not.

Davidski said...

@Chad

Hmm.. Not sure how the authors missed this...

I'm getting much better fits with Neolithic and Copper Age samples from Europe.

The Maykop samples aren't necessary to fit Yamnaya, so what's the point?

Chad said...

Here's that three-way..

left pops:
Steppe_EMBA
Ukraine_N
Maykop_Novosvobodnaya
Progress_Eneolithic

right pops:
Chimp
Iron_Gates
Ust_Ishim
EHG
West_Siberia_N
Levant_N
Ganj_Dareh_N
Anatolia_N
Brazil_LopaDoSanto_9600BP
CHG

numsnps used: 568680

best coefficients: 0.163 0.130 0.707
std. errors: 0.015 0.021 0.030

fixed pat wt dof chisq tail prob
000 0 7 8.891 0.260588 0.163 0.130 0.707

Chad said...

David,

Whether fits have a tail probability of .6 or .3 doesn't really matter. Both are successful. If one was 9.879e-05, that's another story. The thing is that mine has other backing,\; that being archeology. Is there Maykop influence? Yes. Do we have many more examples of influence involving geneflow or not? Very overwhelmingly, aDNA is showing that cultural changes almost always involve some change to the DNA. Is it plausible that Repin is like Progress plus Ukraine N? Could be, considering the geography. I think this is very plausible.

Chad said...

If we had Repin, we may be able to make Yamnaya a mix of Repin and Maykop. These Steppe EMBA are just Samara Valley samples. Yamnaya Caucasus has a good chunk of Maykop.

Davidski said...

@Matt

@Chad, probably worth to message the authors on that one if it fits as well, esp. if can accommodate all same pops as their final 3 pop model with similar p value.

Better not, unless he can demonstrate with different sets of outgroups that the Maykop pops are a better alternative to European farmers from Ukraine, Hungary etc.

When I compare the Caucasus and Europe as mixture sources for Yamnaya, the Caucasus is generally rejected, like in that model above for Sredny Stog, in which Meshoko is given a negative coefficient.


Chad said...

Matt,

I did email Dr. Haak the outputs

Davidski said...

@Chad

It's not just about comparing the tail probs in a couple of models. My analyses consistently show better fits with European farmers than with Maykop.

And of course this is in line with archeological data which show that Yamnaya is a mixture of Repin, Khvalynsk and Sredny Stog, not Maykop.

So I'm not sure why you think Maykop is a plausible source of significant ancestry in Yamnaya?

Another issue that you have is that Yamnaya shows different Y-DNA and mtDNA from Maykop, while it shares mtDNA with European farmer groups.

I just don't see the point of what you're trying to achieve? Can you run a detailed analysis comparing European farmers and Maykop as sources of Anatolian ancestry in Yamnaya, and post a whole series of tail probs so we can look for patterns? Like here...

Yamnaya: home-grown



andrew said...

@Azra The abstract to the Koch paper that you reference includes this statement: "The authors accordingly suggested that the Rathlin men spoke the Indo-European language that then evolved into Gaelic in situ."

This is disappointing, because that is one of the outstanding intra-Indo-European linguistic issues least definitively resolved by ancient DNA.

The geographic distribution of the Celtic languages, their relative similarity to each other (reflective of a fairly shallow time depth) and the material culture of the earliest historically attested Celtic speakers, all tend to favor the alternative conclusion that even if the Rathlin men spoke an Indo-European language that the languages derived from proto-Celtic, including Gaelic arrived later.

The great genetic upheaval in that area upon the arrival of the Bell Beaker people absolutely strongly favors a language shift occurring at that time (probably from a non-Indo-European Neolithic language in the same language family as that of other Anatolian derived first farmers of Europe). But, just what languages was shifted to is muddier than ever.

The genetics definitely makes the hypothesis that the Celtic languages spread with the Urnfield culture in the early Iron Age much less definitive. But, it also doesn't provide very strong affirmative evidence for an in situ development of Gaelic without any intermediate languages being spoken. And, in the absence of any writing or even any sufficiently specific oral histories from that era, this question may never be possible to resolve authoritatively.

Samuel Andrews said...

@Chad,
"These Steppe EMBA are just Samara Valley samples. Yamnaya Caucasus has a good chunk of Maykop."

Yamnaya R1b Z2103+ was a very successful clan. The entire Yamnaya horizon was mostly uniform genetically. Wherever, they went they typically kept to themselves. Yamnaya Caucasus has little or no Maykop ancestry.


Afanasievo in Siberia was basically 100% Yamnaya (Repin whatever). Yamnaya in Caucasus was also basically 100% Yamnaya.

You should be aware of this trend by now.

3.0866"

Yamnaya_Caucasus

Progress_Eneolithic:PG2004,87.4
Globular_Amphora_Ukraine:ILK002, ILK003,5.2
Globular_Amphora_Poland:I2403,3.6
Balkans_ChL,2.9
Darkveti-Meshoko,0.9

Davidski said...

@Chad

Here's another pair of models that show how Caucasus ancestry is outright rejected on the steppe when pitted directly against a more western source of Anatolian-related ancestry.

It was Sredny Stog before, and now here's Yamnaya Caucasus, which, according to you, has a decent chunk of Maykop ancestry.

The fit in the successful model isn't great, but that's not the point. The point is that a stream of ancestry from the Caucasus breaks the model.

Yamnaya_Caucasus
Maykop_Novosvobodnaya -0.045±0.118
Progress_Eneolithic 0.803±0.095
Ukraine_N 0.067±0.048
Varna 0.175±0.072
chisq 14.662
tail prob 0.144878
Full output (infeasible)

Yamnaya_Caucasus
Progress_Eneolithic 0.772±0.030
Ukraine_N 0.079±0.024
Varna 0.149±0.021
chisq 18.542
tail prob 0.0698188
Full output

a said...

Why on earth would Maykop give Yamnaya the tools for colonization? If they had there own metal/pottery/wagons/oxen/equipment-they could have easily colonized the vast open steppe without the help of Yamnaya,unless of course they were not familiar with the climate/terrain and or hostile violent/clans. So why did there attempt at colonization fail?

a said...



https://www.youtube.com/watch?v=L0aFqzKTBp8

Shishlina 2018 - Genes, Isotopes and Artefacts Conference Vienna

Davidski said...

Yeah, the attempted colonization of the steppe by Maykop failed.

And Yamnaya doesn't have any Maykop ancestry, because as per archeological data, it's a mix of Repin, Khvalynsk and Sredny Stog.

Not sure what Chad is pushing here and why.

a said...

Something also interesting is R-V1636 uniparental. Found in Khvalynsk Progress(2)and Armenia.With the oldest sample from Khvalynsk burial R1b -R1a and (Q)contain the oldest dated copper samples in Volga region. The authors input possible origin of copper with trace elements from south east Europe.

Davidski said...

@All

By the way, if anyone's worried that the data might not be reliable, I tried reproducing various models and stats from the Wang et al. paper and they all worked very well.

Here's a qpAdm model for two (out of the four) Yamnaya_Caucasus samples from the paper using most of the same outgroups as in the paper (I don't have most of those moderns marked with DG), and despite this it's very similar to that in the paper.

Yamnaya_Caucasus
Barcin_N 0.222±0.046
Eneolithic_steppe 0.742±0.045
WHG 0.035±0.019
chisq 5.928
tail prob 0.655326
Full output

Drago said...

@ Matt

''7000 BP (5000 BCE) is, indeed, approximately 2000 years before the Steppe_Maykop samples. If there were, for some reason, reason to be confident that the expansion of CHG ancestry north of the Caucasus had to wait until the dawn of the neolithic in the Caucasus. Which there is not.''

I'm not sure why we're misunderstanding each other here. As I suggested, lets focus on the region at hand. The issue is that Mesolithic / Neolithic settlement in the north Caucasus is rather sparse, save a rock shelter here and there; and this contrasts with the plentifully inhabited Dnieper rapids, or West Georgia, by sub-Neolithic hunter-gatherers. This all begins to change only c. 4500 BC. The earliest set of samples we have seem to have, at best, only trace amounts of Siberian_N ancestry, then it profoundly rises in the Steppe-Majkop individuals (but a few hundred years later) which follow Piedmont Eneolithic, and the region generally sees a furthwre growth in sites.

So the problem (at present) isn't CHG (as nearby Georgia +/- the rock chelters, lower Don, Crimea could all account for the long-presence of CHG); but Siberian N ancestry is the mystery, at least to me. Thus, the economical explanation as currently stands is that this population came from somewhere and with someone- not necessarily exactly in 4500 BC, but still within a Neolithic time frame.

Drago said...

This is useful from the Supp, for those still familiarizing

''The earliest attested evidence of the Neolithic lifestyle in the North Caucasus, including
domesticates and settlement architecture, dates to the mid-5th millennium BCE and is associated with a cultural formation termed Darkveti-Meshoko Eneolithic or ‘pearl-ornamented ceramic’. ''

It's important to be cognizant that whilst the 5th Mill is a ''Copper Age'' in East Balkans, or Near East, in places like the Caucasus, the steppe, or Brittany, its Neolithic. In others (eg Baltic), still Mesolithic.

Davidski said...

In practice, the 5th millennium BCE is actually still the Mesolithic on the steppe just north of the Caucasus.

And that's why the CHG in Yamnaya isn't from any farmers, but from steppe foragers.

Drago said...

Yeah maybe in terms of lacking agriculture or domesticated animals. But sociologically, it certain wasn’t Mesolithic. The impact of interacting with groups like Varna and Haemangia profoundly altered social structure- but this played out differently in different subregion of the steppe.
I’m beginning to think that Progress etc are new colonists arriving from the north; but acquiring significant CHG in the process

Open Genomes said...

Some unexpected migrations are now being revealed by the new samples, and these fill in the blanks for existing samples we've seen before.

These new samples hold some surprises when we look at some of them on the Global25 Ward's distance-squared clustering tree:

Eurogenes Global25 Ward's distance-squared clustering tree

Some closeups of the tree worth focusing on, before I show the restricted nMonte analysis for individual samples:

Steppe Maykop on the Global25 Ward's distance-squared clustering tree

The Steppe Maykop individuals cluster with Botai / West Siberian Neolithic, and the various Andronovo, Sintashta, and Srubnaya "outliers" who also derive from Botai / West Siberian ancestry.

Progress and Vonyuchka Eneolithic and Steppe Maykop AY2001 on the Global25 Ward's distance-squared clustering tree

Progress and Vonyuchka Eneolithic, and Steppe Maykop AY2001 form a divergent steppe clade with the I1783 Gonur1 BA outler and I4784 Kyzlbulak Middle Late Bronze Age II, which is the result of admixture.

Maykop Novosvobodnaya on the Global25 Ward's distance-squared clustering tree

Maykop Novosvobodnaya clusters with Dzharkutan 2 Bronze Age female outlier, who is now proven to clearly be a Maykop woman living all the way out in BMAC, just north of Afghanistan. They cluster mostly with Abkhazians. It seems likely that Maykop Novosvobodnaya were in fact Northwest Caucasians, and the presence of J-Y11200 bears this out.

Kura-Araxes on the Global25 Ward's distance-squared clustering tree

On the other hand, the Kura-Araxes individuals cluster with present-day Kartvelians. It seems quite likely that Kura-Araxes were proto-Kartvelian speakers, and perhaps with the individual from Velikent in J-Z1842 who is a bit different from the others might have been a proto-Northeast Caucasian.

Individual nMonte analyses to follow.

Ric Hern said...

When did Lake Baikal related Pottery enter Europe ? The distinction between true Neolithic(Pottery+Domesticates) and Pottery Neolithic (Late Upper Paleolithic/Mesolithic) Pottery seems to confuse some people...

Davidski said...

@Chad

This is about the only decent model that I can find for Yamnaya_Caucasus with my outgroups, which, admittedly, are tough to work with, but that's a good thing.

Yamnaya_Caucasus
Progress_Eneolithic 0.780±0.027
Trypillia 0.148±0.022
Ukraine_N 0.072±0.024
chisq 16.577
tail prob 0.121023
Full output

Adding Maykop_Late doesn't improve the model. And adding Maykop_Novosvobodnaya and Steppe_Maykop is just pointless.

I suppose that Yamnaya_Caucasus might have some minor late Maykop ancestry that it picked up as Maykop was disintegrating, but even that seems a long shot to be honest. Haak probably knows this already though, considering what was in his paper.

Yamnaya_Caucasus
Maykop_Late 0.107±0.130
Progress_Eneolithic 0.704±0.098
Trypillia 0.080±0.085
Ukraine_N 0.109±0.051
chisq 15.436
tail prob 0.116962
Full output

Nothing to see...

Maykop_Novosvobodnaya

Steppe_Maykop

Mem said...

Kura araxes were clearly hurro-urartian speakers.Except in northern Iran and Northeast Caucasia,those were populated by Nakh-dagestan speakers.

Kartvelian languages spoken in kolkhis plain and svaneti at bronze age only during iron age they were expand through Caucasia.

Davidski said...

@Mem

I'm certain that Shulaveri-Shomu samples will be practically identical to those of Kura-Araxes Kaps, with lots of Y-hg J and G, and no R1b-M269.

So what language do you think Shulaveri-Shomu spoke?

Matt said...

@Chad:
Is it plausible that Repin is like Progress plus Ukraine N? Could be, considering the geography. I think this is very plausible.

If we had Repin, we may be able to make Yamnaya a mix of Repin and Maykop. These Steppe EMBA are just Samara Valley samples. Yamnaya Caucasus has a good chunk of Maykop.

Nothing obviously implausible there at all as far as I know.

My hope is that there is enough skeletal material to obtain a good volume of samples from the relevant cultures in the relevant time frame, and some of them will only be able to modelled one way or another.

Then we can dispense with a lot of discussion about which different qpAdm models out of multiple plausible models work best (which outgroups, which p-value) and best guesses on genetic affiliation from archaeological affiliation (which is more informative than I have often thought yes, but still uncertain as cultural influence and genetic admixture can become unlinked).

(Again this study didn't show any interaction between Maykop and populations near the northern coast of Black Sea that could be plausibly ancestral to Yamnaya... but they didn't have any samples in the right kind of contact zones, only contact zones with circum-Caspian, which seems CHG+EHG+West_Siberian, and distinct.)

Repin Culture is pretty vital here. Esp. in light of speculation that Afanasievo was an offshoot of Repin Culture before formation of Yamnaya. More Khvalysnk with solid dating as well. Etc.

Davidski said...

@Matt

Nothing obviously implausible there at all as far as I know.

Pretty sure this is...

Yamnaya Caucasus has a good chunk of Maykop.

Slumbery said...

@Chad Rohlfsen
"These Steppe EMBA are just Samara Valley samples."

Well, Yamnaya Kalmykia is closer to the Caucasus than to the Samara river.

G25 nMontes Yamnaya_Samara:Average
Khvalynsk_Eneolithic 31.67%
Maykop_Late 5%
Progress_Eneolithic 55.33%
Trypillia 5%
Fit 3.4965

G25 nMontes Yamnaya_Kalmykia:Average
Khvalynsk_Eneolithic 29.17%
Maykop_Late 10%
Progress_Eneolithic 55.83%
Trypillia 5%
Fit 3.2483

G25 nMontes Yamnaya_Caucasus:Average
Khvalynsk_Eneolithic 21.67%
Maykop_Late 10.83%
Progress_Eneolithic 64.17%
Trypillia 3.33%
Fit 3.1119

@Matt
" in light of speculation that Afanasievo was an offshoot of Repin Culture before formation of Yamnaya."

Afanasievo is very close to Yamnaya genetically, especially to Yamnaya_Kalmykia, so this would also assume that Repin was essentially identical to Yamnaya genetically and there is a straight ancestor-descendant relationship between them with not much else coming in. Maybe so, I do not know.

Matt said...

@Davidski, yeah, I'm not sure about that, because it depends on models. The bit I was looking to say was not implausible is: "If we had Repin, we may be able to make Yamnaya a mix of Repin and Maykop.", which is plausible until samples clarify one way or the other. Quoted a bit too much of him there.

Matt said...

@Slumbery, yeah, this is my reaction to that as well. I'd actually expect samples from Repin to be more likely to disconfirm Anthony's Afanasievo=direct offshoot of Repin conjecture, but samples would clear it up one way or the other.

Samuel Andrews said...

Sredney Stog R1a M417+ guy picks Globular & SwedenEN as his farmer ancestors. But this could not possibly be correct. RIght? Because, he existed before Globular Amphora. His only possible farmer ancestor is Cucuteni–Trypillia. Right? They were the only farmers who settled east enough to be in direct contact with early Proto Indo Europeans.

His excess HG ancestry must be from UkraineHGs. This would mean Yamnaya have a little bit of UkraineHG. Corded Ware-derived Europeans, Bell Beaker & Andronovo, might have more.

Globular Amphora & SwedenEN are both about 10% EHG (aka 5-7% ANE). They have a decent amount of Romania/UkraineHG-like ancestry that their western/southern European relatives lack. This is why Sredny Stog chooses them.


@OpenGenomes,
"Maykop Novosvobodnaya clusters with Dzharkutan 2 Bronze Age female outlier, who is now proven to clearly be a Maykop woman living all the way out in BMAC, just north of Afghanistan."

That's fascinating.

Samuel Andrews said...

Varna_O also has a lot of HG ancestry. She picks Sredny Stog as her Steppe ancestor.

Slumbery said...

@Samuel Andrews
His only possible farmer ancestor is Cucuteni–Trypillia. Right? They were the only farmers who settled east enough to be in direct contact with early Proto Indo Europeans.

LBK advanced as far east as the Vistula basin and that brought it and its local successors pretty close to the later Dnieper-Donets. Vistula LBK and the following local farmer communities are unsampled.
Also Funnelbeaker. It was all over Poland down to Ukraine. That would also explain the detection of Swedish Neolithic ancestry portion.


Mem said...

Shulaveri culture is very ancient.circa 5000-4000 BC maybe even older.So,I think they spoke a ancestral language for all kartvelian,NWC,NEC,Hurri-Urarti.

Andrzejewski said...

@Mem Not likely. Because these languages are very different from each other.

Slumbery said...

@JuanRivera
"Ran some models of steppe cultures using Maykpo. For all cultures, some Maykop input is visible, but always being minor. In Sredny Stog, the model chose Maykop_Novosvobodnaya over both european farmers and Progress_Eneolithic, but the fits were worse than the models without it."

What kind of test did you run? I ask that, because I just tried G25 nMonte on both the Ukraine_Eneolithic:average and specifically Ukraine_Eneolithic:I6561 and both either rejected Maykop_Novosvobodnaya in the presence of European farmers and Progress_Eneolothic or at least it remained a minor component compared to them (depending on the exact reference populations).



Samuel Andrews said...

JuanRivera, try modelling Steppe_Eneolithic with only pure stone age sources: Barcin_N, LevantN, EHG, Afontova-Gora, CHG, Ganji. See if they score Barcin_N. If they don't that probably means their Maykop scores are noise since Maykop had a lot of BarcinN/LevantN ancestry.

@All,

I'd like to see Reich, Lazaridis, others response to the news that a Caucasus/Near East population didn't give rise to Yamnaya & therefore it is likely Proto Indo European is native to the Russian Steppe.

Slumbery said...


"try modelling Steppe_Eneolithic with only pure stone age sources: Barcin_N, LevantN, EHG, Afontova-Gora, CHG, Ganji. See if they score Barcin_N."

Ukraine_Eneolthic would surely score Barcin_N with those references, because it has European farmer ancestry and from those references only Barcin_N can stand for that. But that does not mean that the Maykop signal is real.

Some example runs:

Global 25 nMontes Ukraine_Eneolithic:Average
Comb_Ceramic 46.67%
Globular_Amphora_Poland 32.5%
Maykop_Novosvobodnaya 0%
Progress_Eneolithic 20.83%
Fit 2.2633

(One could argue that Globular Amfora is younger than parts of Ukraine_Eneolithic, but so is Maykop and GB does not have mixture from CHG/Caucasus, not even indirectly, so it is a good reference for a WHG rich farmer source.)

Global 25 nMontes Ukraine_Eneolithic:I6561
Comb_Ceramic 6.67%
Globular_Amphora_Poland 36.67%
Maykop_Novosvobodnaya 0%
Progress_Eneolithic 56.67%
Fit 4.0896

The fit is much worse for this rather outlying sample, but the point is: Maykop is still rejected.

And to try your suggestion:

Global 25 nMontes Ukraine_Eneolithic:Average
Barcin_N 29.17%
CHG 6.67%
EHG 45.83%
WHG 18.33%
Fit 3.8178

Global 25 nMontes Ukraine_Eneolithic:I6561
Barcin_N 32.5%
CHG 15%
EHG 52.5%
WHG 0%
Fit 5.6421

Sure enough, a lot of Anatolian there, just not from Maykop. Also however this makes me think that the farmer ancestry of I6561 was not so WHG rich after all. (Also note that nMontes always underestimate CHG in the presence of EHG, but that is another story.)


So let us see the more CHG rich I6561 with a less WHG farmer source and since it did not react as well to Comb Ceramic as the average, some other EHG-rich source from the steppe.

Global 25 nMontes Ukraine_Eneolithic:I6561
Khvalynsk_Eneolithic 38.33%
Maykop_Novosvobodnaya 9.17%
Progress_Eneolithic 19.17%
Trypillia 33.33%
Fit 4.3657

Now we have some Maykop, but it definitely not preferred over European farmers and Progress.

Net us see CHG vs. Iran_N in your suggested deep ancestry.

Global 25 nMontes Ukraine_Eneolithic:I6561
Barcin_N 33.33%
CHG 12.5%
EHG 51.67%
Ganj_Dareh_N 2.5%
Fit 5.6782

It takes some Iran_N, but I would not read too much into it.

This is not the full picture, but I'd say I6561 does not have a big ancestry from beyond the Caucasus. Maybe some.




Chad said...

I'll get to the graphs and such with the new data, but it will have to wait until I finish working on my new blog post. It could take a few days.

Andrzejewski said...

@Samuel Andrews "
'd like to see Reich, Lazaridis, others response to the news that a Caucasus/Near East population didn't give rise to Yamnaya & therefore it is likely Proto Indo European is native to the Russian Steppe."

I've just emailed it to Dr. Lazaridis.






Davidski said...

@Matt

The bit I was looking to say was not implausible is: "If we had Repin, we may be able to make Yamnaya a mix of Repin and Maykop.", which is plausible until samples clarify one way or the other.

You and Chad seem to be forgetting that Yamnaya, including Yamnaya Caucasus, has minimal Anatolian input. In fact, it's so small that it must be heavily diluted with WHG to make sense of all of the results from various types of analyses, including consistent results with formal stats, ADMIXTURE and PCA. And I can add uniparental marker data to this.

So claiming that Yamnaya possibly has significant Maykop ancestry doesn't gel with this fact, and no, it isn't plausible.

Andrzejewski said...

@Davidski The Cimmerians used to be known as related to the Scythians and Sarmatians. However it seems like they carry lots of East Eurasian admixture:

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

Would you recognize Cimmerians as stemming from the Karasuk Culture in the Eastern Steppes?

Folker said...

In any case, Wang et al. have tested Eneolithic Caucasus for various Steppe populations (supp table 16), and this was rejected in most of the cases (except for Yamnaya Caucasus and Yamnaya Ukraine Ozera). Since they modeled Maykop as mostly Eneolithic Caucasus, it is explaining why they didn't push more complex scenari, as the prerequisites are not met.
By the way, as already said, uniparental markers are completely different between Caucasus and Steppe Eneolithic and Yamnaya (if you put aside Steppe Maykop, and the Ozera sample + Yamnaya Caucasus). It is therefore clear that Yamnaya didn't expanded from the Caucasus, but are intrusive there, and Yamnaya Caucasus are representative of an expansion of the Yamnaya Horizon. The Ozera sample has clearly a Caucasus mother, but is a clear outlier.
They are using the "tracer dye" theory because they have failed to find any real Maykop admixture in Yamnaya. From their own words.

Davidski said...

@Andrzejewski

Would you recognize Cimmerians as stemming from the Karasuk Culture in the Eastern Steppes?

Probably, or from a closely related group, but keep in mind also that some Scythian groups had more East Asian ancestry than Cimmerians.

Slumbery said...

@Andrzejewski

Actually Karasuk (mostly like Karasuk outlier actually) is better fit for the Saka groups than to Cimmerians. Cimmerians are not Eastern at all compared to Asian Saka&Scytian groups. They are eastern compared to Scytian_Moldova and Scytian_Ukraine sure, but those are much more western than other Scytians, except Scytian_Hungary (the latter is almost fully local/regional genetically).

So the Cimmerians can be pretty much related to Saka&Scytians, but not the local ones, the Asian ones. (The regional Scytian samples are all younger than the Cimmerian ones anyway.)


Also we have Arabic spammers now, this is a milestone in the history of the blog. :)

Matt said...

@Davidski, well, the reason Chad's models work is going to be because a population admixed between Maykop+Ukraine_N and then admixed with Piedmont_Eneolithic, can accommodate Yamnaya. The WHG:AN ratio can be supplied by Ukraine_N. 2-way Maykop+ UkraineN or EHG do not seem like viable models, but that does.

I don't think we have enough samples to disregard that possiblity, and there is a Maykop presence in the western region north of the Caucasus and a broader "Maykop sphere of influence" that is unsampled, and whether the influence coincided with any ancestry.

Visually: https://imgur.com/a/8mxXEq9

That's why my hope is that they will be able to sample that region and the relevant candidate predecessor cultures for Yamnaya and be able to properly eliminate one or the other based on the time and positions of samples.
E.g. "Nope, too much farmer admixture at this time for that to work. No Ukraine_Neolithic types left" or "Nope. No samples really fit a cline between Ukraine_N and Maykop; only Trypillia and Ukraine N".

Davidski said...

@Matt

You need to look at all of the evidence, and keep in mind that the analyses based on formal stats that Chad is relying so heavily on can produce biased results depending on how they're set up.

But I'll explain everything soon.

Synome said...

@Matt

Can you point me to some academic work regarding the "Maykop sphere of influence" that is currently unsampled? Is that in the Caucasus paper or somewhere else?

Matt said...

@Synome, yep, Caucasus paper designates a Maykop "sphere of influence" and a direct Maykop zone that includes regions of the mountain flank at the west that is not sampled. Going off that, so perhaps other archaeologists would deny that there ever was a Maykop "sphere of influence".

@Davidski, thanks for adding the other new samples.

Sticking them with the same base West Eurasian ancient "scaffold" in G25 as I used in my post above, then reprocessing: https://imgur.com/a/S9AzeRW

Lola does look to me more intermediate the main Steppe_Maykop cluster (which I class as SA6004, AY2003 and SA6001) and Catacomba/Yamnaya/Kubano-Tersk, and distinct from the Steppe_Maykop outliers AY2001, SA6012, IV3002. So that looks to me to support Lola being a mixture of a pop like main Steppe_Maykop and Steppe_EMBA like groups (about 40:60 or 50:50) and not really a continuation of the Steppe_Maykop outliers. I would conjecture that groups like main Steppe_Maykop persisted in arid areas, as re-appearance with Lola is linked to aridification of climate?

IV3002 (most western steppe maykop outlier) looks off cline for being purely intermediate the Caucasus cluster and the main S_Maykop cluster and looks as if it needs some from Piedmont_Eneolithic or something related to Yamnaya as well, but this is hard to call.

Within Caucasus cluster proper, with more samples, it looks more robust that Darkveti-Meshoko, NC_MBA and Maykop_Novosvobodnaya have a bit more genuine CHG ancestry, while the KA samples (esp. KA_Talin) have more Seh_Gabi_Chl+Barcin/Levant related ancestry that is only pseudo-CHG in some dimensions. Velikent Kura-Araxes VEK007-009 looks a little EHG/Steppe_Maykop shifted (compatible with Velikent in east and being off the mountain flank?). Maykop OSS001 looks both EHG and Barcin shifted? But these are hard to tell differences.

Matt said...

Putting the Sintashta on the scaffold reveals that Lola is quite similar to Sintashta outlier I0983 from the Sintashta outlier o1 cluster, which as Slumbery rightly explained appears to lack admixture from main Sintashta cluster but is between West_Siberia_N and Yamnaya:

https://imgur.com/a/lBhMbuL

Though I'd guess that from Sintashta outlier o1 cluster, only I0983 is truly perhaps between Yamnaya and West_Siberia_N. I1017 and I1007 look to have excess East Asian ancestry which, can't be explained by West_Siberia_N (unlike Lola:NV3001 or I0983).

Davidski, can I also suggest that Sintashta:I7670 is broken out in your dataset from the main Sintashta in G25?

He/she seems to plot further from the main Sintashta cluster than Sintashta_o2 does, and basically looks like a Steppe_EMBA with a dose of extra EHG.

See: https://i.imgur.com/8usNI1Q.png

Davidski said...

Yeah, that's a wrongly labeled Potapovka sample. Forgot to fix that in the G25 datasheets.

Andrzejewski said...

@Davidski @Slumbery "The authors found it "noteworthy that the oldest of the Cimmerians studied here (cim357) carried almost equal proportions of Asian and West Eurasian components, resembling the Pazyryks, Aldy-Bel, and Iron Age individuals from Russia and Kazakhstan. The second oldest Cimmerian (cim358) was also the only one with both uniparental markers pointing toward East Asia. The Q1* Y chromosome sublineage of Q-M242..."

The importance is that Cimmerians used to be thought of as descending from the Srubna Culture but now we see that they are more of an East Asia (actually, Ancient Northern Eurasian) intrusion into Steppe Indo-Iranic groups.

Also, while Scythians/Saka and Sarmatians predominantly carry the R1a1 R1a-M269 paternal signal and both Western Eurasian and Eastern Eurasian mtDNA, with Cimmerians carrying Q-M242 betrays an unusual pattern.

I also recall reading that the Poltavka and heirs populations all the way to Scythio-Sarmatians used to be described as having strong Europoid features by Anthropologists, despite all the admixture events leading there: Poltavka mixed with non-IE non-Uralic Volsovo, then Strubna mixed with Botail or other W Siberians, then Andronovo with BMAC (part W. Siberian, part Iran_Chl), in addition to Afanasievo mixing extensively with the Yenisian/Botai Okunevo to create Karasuk and then Tagar: all these Steppe-derived populations maintained the Steppe uniparental marker R1a1-M269 and the pronounced Europoid appearance; but here in light of recent scientific papers we see a shift to an ANE-shifted population with the Cimmerians having paternal ANE haplogroups Q1-M242 but also the whole aDNA is 50% European/Steppe, 50% Asian.

This is significant!

a said...

The worst nightmare seems to be unfolding for those who did not think R1a/R1b people had anything to do with the spread of new ideas and or language; without the direct or indirect help/influence of middle east proper cultures. With all the luck[so far- every published sample] going in the direction of R1a/R1b steppe sphere of influence. In contrast, using various lines of creative reasoning like no samples from philosophical phantom spheres of influence + low but detectable creative tracer dye movements of either genes or culture + creative middle east biased monikers , [all of which can be used in an opposite line of reasoning-direction]. Again, as luck[co-incidence] would have it, the genetic spread- influence from steppe to middle east and other regions connected to Europe, has already been conceded with genetic samples from [26kya]- Dzudzuana Georgia samples showing a north to south movement. It will be interesting to see more genetic samples released; like the ones dealing with wagon/wheel building cultures-animal husbandry in the use of oxen/horse for sedentary cultural expansion versus mobile cultural expansion, in the Caucasus and steppe.

Matt said...

@Davidski; Ah, cool. That adds to the perception of diversity in the Potapovka samples compared to the Sintashta samples, and the Srubnaya culture that succeeds Potapovka (which is basically like Sintashta).

Of course Sintashta has diversity, but there are 29 main samples that are mainly like each other and 8 outliers, while the Potapovka has 4 samples which are all quite different from each other - 1 like Sintashta main cluster (I0419 -M: R1a1a1b2, U2e1), 1 fits in with Poltavka (I7489 - F: H2a1e), 1 like Poltavka with more EHG (I7670 - M: R1b1a1a2, J1c1b1a), and 1 like Lola:NV3001/Sintashta outlier:I0983/Srubnaya outlier:I0354 (I0244 - F, U2e1a1).

Obviously I7670 with R1b1a1a2 looks like male lineally Poltavka, while I0419 is male lineally related to the Corded Ware R1a types (like Sintashta and the succeeding Srubnaya).

Samuel Andrews said...

Beaker Rhine & Sintashta/Andrnovo are two ethnic groups that emerged at western & eastern edges of Corded Ware culture.

Slumbery said...

@Andrzejewski

I only said that the Cimmerians can be very well related to Asian Saka&Scytian, because they are not particularly eastern or even that Heaven vs. Earth different compared to them. There is no doubt that Cimmerians are very eastern for their Geographical position.

Even your source note that: "The authors found it "noteworthy that the oldest of the Cimmerians studied here (cim357) carried almost equal proportions of Asian and West Eurasian components, resembling the Pazyryks, Aldy-Bel,..."

Those are groups dubbed as Scytian in the Global 25 datasheet.

"The importance is that Cimmerians used to be thought of as descending from the Srubna Culture but now we see that they are more of an East Asia (actually, Ancient Northern Eurasian) intrusion into Steppe Indo-Iranic groups."

They are definitely not Srubnaya descendants genetically, I doubt they even have as much as a double digit ancestry percentage from them. However the same could be said about the Asian Saka&Scytians. Sarmatians are another story.

"...despite all the admixture events leading there: Poltavka mixed with non-IE non-Uralic Volsovo, then Strubna mixed with Botail or other W Siberians, then Andronovo with BMAC (part W. Siberian, part Iran_Chl),..."

You draw a line of descent here that did not really exist in this way.
1. Poltavka does not have significant Volosovo admixture. At least not enough to make them any more EHG or Siberian than their local predecessor, Yamnaya Samara was.
2. The Srubna samples we have do not show Siberian admixture. In fact they are very western, significantly more western than Poltavka and have only partial decent from Poltavka.
3. Andronovo's European side seem to descend mainly from Sintashta-Petrovka, not from Srubnaya.

...in addition to Afanasievo mixing extensively with the Yenisian/Botai Okunevo to create Karasuk...

Karasuk had significantly more Anatolian Neolithic ancestry that Afanasievo had, so I'd add Sintashta-Petrovka to that mix.

...but here in light of recent scientific papers we see a shift to an ANE-shifted population with the Cimmerians having paternal ANE haplogroups Q1-M242 but also the whole aDNA is 50% European/Steppe, 50% Asian."

I'd estimate that the average of the G25 Cimmerians had maybe 30% Siberian/North-Central Asian ancestry in Mezo-Neolithic time dept and inside that their WSHG/Botai ancestry is less that of the Sarmatians (who are more western overall, because of less BHG&East Asian ancestry.) Actually Cimmerians even had some East Asian proper ancestry and they can be modelled in nMonte with zero Botai/WSHG.

For example:

Global25 nMontes Cimmerian_Moldova:Average
Afanasievo 9.17%
Dzharkutan2_BA 20.83%
Shamanka_BA 12.5%
Sintashta_MLBA 50%
West_Siberia_N 0%
Xiong_Nu 7.5%
Fit 3.0957

The Cimmerian samples are extremely variable in their ancestry proportions, it was apparently a very heterogeneous population, but ironically the one mentioned by the source, "cim357" has the least Siberian+East Asian ancestry from the three Cimmerian in the Global 25 datasheet. Definitely nowhere near 50% Asian in the sense used there.

I do not think the Cimmerians are a good proxy to see how the ANE people looked like.

Andrzejewski said...

@Davidski Would the new findings about Cimmerians having ANE/East Asian ancestry rule out them (and some Scythian/Saka, maybe even Tocharian groups) being Indo-European speaking?

Matt said...

Had a go at some nMonte fits for the Steppe_Maykop samples that are outlying (SA6013, IV3002, AY2001), and for Lola:NV3001.

Full dump of nMonte here: https://pastebin.com/KEhyX3tZ

Now the first thing I found was that with the most unrestricted set, there wasn't much pushing the samples to take the proximate likely samples from the region: Steppe_Maykop and the Caucasus samples. Instead tended to take deeper samples and had no preference for locals over Central Asian Eneolithic, and also no clear preference for taking Maykop over CHG+EEF and so on.

Eventually after using the most restrictive set of only those that were most proximate in the region at the time I did find the following:

* SA6013: 41% Steppe_Maykop main cluster, 51% Maykop(Maykop+LateMaykop+Novosvobodnaya), 7% Vonyuchka_Eneo, 0.6% Ukraine_N

* IV3002: 61% Maykop (Maykop+LateMaykop+Novosvobodnaya), 26% Progress Eneolithic, 9% Steppe_Maykop main cluster, 4% EHG

* AY2001: 66.6% Steppe_Maykop main cluster, 29.6% Maykop (Maykop+LateMaykop+Novosvobodnaya), 3.8% Vonyuchka_Eneolithic

* Lola (NV3001): 62.8% Steppe_Maykop main cluster, 27.2% Progress_Eneolithic, 4.4% Maykop-Late, 4.4%, Ukraine_N, 1.2% EHG.

(The last fit for Lola is probably an abstraction of Steppe_Maykop+Catacomb / Kubano-Tersk).

So not 100% confident with this, but that fits with my first impressions that 1) SA6013+AY2001 are clinal between the main Steppe_Maykop group and Maykop, 2) IV3002 is mostly much more clinal between something like Progress_Eneolithic and Maykop, with less of the Siberian influence and 3) Lola is roughly intermediate Steppe_EMBA and Steppe_Maykop main cluster, not Caucasus and Steppe_Maykop main cluster.

Andrzejewski said...

@Slumbery Will all these findings mean that Cimmerians or even Scythians plus some Yuezhi/Tocharians may NOT be IE speakers after all?

Now, some people would regard Tocharians as descendants of the Afanasievo. Here is why they are wrong:

1. Afanasievo mixed with Okunevo, which may be related to Yeniseian/Botai/WSHG. The following Prewarski/Tagar Cultures were thought to be Scythian related but maybe they were "Eurasian": part European part East Asia. We know that most of the samples found by "Scythian"/Saka related cultures were R1a1 dominated. However, Tarim Basin Mummies were exclusively so far R1b.

2. Yuezhi/Kuchan/Tocharian was a Centum language, while Saka spoke Indo-Iranian Satem languages.

3. There was a 2,000 year gap between Afanasievo and Tocharian proper.

4. aDNA determined that Tarim Basin Mummies were more closely related to Andronovo than to Afanasievo.

Which means that maybe there was an extinct hitherto-unknown branch of IE that was unrelated to Tocharian. Just like "Belgian" was in between Germanic and Celtic.

5. Xiongnu skeletons (12 of them) were determined to be of Q1a and Q1b Paleo-Siberian ANE rather than Turkic. Moreover, the Jie/Kiat tribe belonging to the Xiongnu confederation fought by Tang Empire were deciphered to have spoken a Yeniseian language rather than a Turkic one. Would you say that the source of Hap Q in Slovak, Hungarian, Romanian and Scandinavian populations (2.5%-5%) came from Huns, Magyars, Avars etc, which may have turned out to be Yeniseian speaking Paleo-Siberians?

6. What do you think happened to the Botai and/or Okunevo after all? Died out? Displaced? Assimilated?

7. Would you agree that the Volsovo were a non-IE non-Uralic Europoid HG (Cro-Magnon?) archaeological culture which may have contributed to the evolution of Balto-Slavic?

Davidski said...

@Andrzejewski

Please try and get the basics right in these sorts of discussions, because it's kind of mind numbing reading stuff like this...

However, Tarim Basin Mummies were exclusively so far R1b.

All but one of the Tarim Basin Mummies belonged to R1a.

Would you agree that the Volsovo were a non-IE non-Uralic Europoid HG (Cro-Magnon?) archaeological culture which may have contributed to the evolution of Balto-Slavic?

Garino-Bor and Volosovo are generally regarded to have been proto-Uralic or early Uralic, and there's no evidence that they had anything to do with the formation of Balto-Slavs or that they looked especially like Cro-Magnons.

Andrzejewski said...

It's amazing how modern scientists are challenging the prevalence of ancient Europeans all over Asia. It was known that around the time of Christ large swaths of land in Asia derived from Steppe populations such as Saka. Wudun, Scythians, Sarmatians, Tocharians/Yuezhi/Kuchans. It was acknowledged until very recently that Central Asia, Anatolia, Northwestern China, Western Mongolia, even parts of Afghanistan-India-Nepal originated in Steppe nomads.

There seems to be a pattern here, to refute the archeological history, whereby in many cases: Kazakhstan, Tarim Basin, Western Mongolia - Europoid populations were the first to arrive.

Now they extend it to the Pazyryk Culture, which was putatively Scythian:

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

"DNA samples recovered from the remains of two Pazyryk males showed them to be members of Y-chromosome haplogroup N1b-P43."

In other words, where did this Uralic/Siberian admixture come from? Scythians used to be exclusively R1a1 on the male side with varying mtDNA markers both West Eurasian and East Eurasian.

Drago said...

Matt
Nice graphs . Wrt the issue of Majkop & Ukraine N as alternatives , can you throw in BA Balkans, cwc , BB. I think they should all line up with “regular”’Yamnaya but if you can it would be nice to visualise

Drago said...

To this graph https://imgur.com/a/8mxXEq9

A said...

@Samuel Andrews

"Sredney Stog R1a M417+ guy picks Globular & SwedenEN as his farmer ancestors. But this could not possibly be correct. RIght? Because, he existed before Globular Amphora. His only possible farmer ancestor is Cucuteni–Trypillia. Right? They were the only farmers who settled east enough to be in direct contact with early Proto Indo Europeans."

Cucuteni-Trypillia + Sredy Stog probably invented wheeled vehicles. The earliest evidence is in Cucuteni-Trypillia.

Samuel Andrews said...

@Andre,
"It's amazing how modern scientists are challenging the prevalence of ancient Europeans all over Asia."

Obviously, mainstream will be sensitive & hesitant towards saying this. But, in 2009 when Andronovo nomads & its descendants' DNA was sequenced the highlight was they had West Eurasian mtDNA & blonde hair and blue eyes. They weren't afraid to say that.

"There seems to be a pattern here, to refute the archeological history, whereby in many cases: Kazakhstan, Tarim Basin, Western Mongolia - Europoid populations were the first to arrive."

It's also obvious you're a chauvinist & Eurocentric. People lived in those places before Andronovo. Also, Iron age Asian Steppe nomads were diverse. All of them had mixed ancestry and None of them were identical to the other. They weren't exactly European/Andronovo but are better described as mixed like modern Mexicans aren't Spanish. Scythian isn't very helpful as a genetic label, it seems all Scythians had mostly different ancestry.

It's a complex story. Turks make it even more complicated.

Andrzejewski said...

@Davidski "Garino-Bor and Volosovo are generally regarded to have been proto-Uralic or early Uralic"

Not according to Jaakko Haakkinen:

https://www.sgr.fi/sust/sust264/sust264_hakkinenj.pdf

He asserts that this Volosovo type substrate which he says is neither IE nor Uralic was the same substrate that Kunda/Narva HG languages were spoken in Lithuania before the advent of Steppe groups.

Matt said...

@Dragos, sure. Here: https://imgur.com/a/pgUML6n

Two slightly different PCA modes used (with basically same results). Gone up to the number of PCs at which populations stop showing tight between pop variation in means and variance.

Slightly busier than the plots you linked as I just stuck the Balkans_BA and Balkans_Chl labelled samples onto the a version I created with more rows.

You'll have to draw the clines on yourself, as I don't have time for that.

If you get the PAST3 software, download this - https://pastebin.com/VWTNfaeW - and rename to .dat file, then you can open in PAST3 and basically follow the instructions here: https://eurogenes.blogspot.com/2018/05/global25-workshop-1-that-classic-west.html, you can replicate all those plots if you want to zoom on them, add extra rows etc. The data is scaled G25 values.

Open Genomes said...

The updated Global25 scaled Ward's distance-squared clustering tree with the additional samples:

Updated Eurogenes Global25 scaled Ward's distance-squared clustering tree

Some highlights:

Steppe Maykop and Lola individuals who cluster with Botai/West Siberia Neolithic and Steppe outliers

Could the "Botai horse tamers" brought horse domestication to the Pontic-Caspian Steppe?

A Steppe Maykop outlier who clusters with Northeast Caucasians

Late Kubano-Tersk, Yamnaya Ukraine outlier, the Hajji Firuz Bronze Age female, and another Steppe Maykop outlier who cluster with Sakas, Cimmerians, and Yaghnobis

Perhaps these individuals are actually Proto-Indo-Iranians?

Progress Eneolithic, Vonyuchka Eneolithic, Yamnaya Caucasus, a Steppe Maykop, and the Gonur Bronze Age outlier, who cluster with Yamnaya, Poltavka, Afanasievo, and some Catacomb individuals

Lake Maykop and a Maykop individual who cluster with Lchashen Metsamor, Armenia Chalcolithic and Armenia Middle Late Bronze Age, a Kumyk, and Azeris from Dagestan

These Maykop / Late Maykop individuals have a more southern affinity than the others.

Kura-Araxes, Darkveti-Meshkoko, Maykop Novosvobodnaya, some Late Maykop, and the Dzharkutan Bronze Age (female) outlier who cluster generally with modern Georgians and Abkhazians and Kotias CHG

Here, in this updated tree, the Kura-Araxes individuals don't favor either Kartvelians or Northwest Caucasians, but rather form a clade with Darkveti-Meshkoko (the North Caucasus predecessor of Maykop), Maykop Novosvobodnaya, and and some Late Maykop individuals. They have general Northwest Caucasian and Kartvelian affinities. The Dzarkutan female outlier from BMAC is clearly a Darkveti-Meskoko individual brought to BMAC, perhaps in trade for tin.

Drago said...

Thanks Matt