search this blog

Friday, April 12, 2019

Armenians vs Georgians


Armenians and Georgians are ethnic groups that live side by side in the south Caucasus, or Transcaucasia. By all accounts, they've both been there since prehistoric times and they're very similar in terms of overall genetic structure.

However, they speak languages from totally unrelated families: Indo-European and Kartvelian, respectively. How did this happen and might the answer lie in the small genetic differences that do exist between them?

To investigate this issue, I ran a series of qpAdm formal mixture models of present-day Armenians and Georgians using tens of ancient reference populations. To come up with as straightforward and meaningful results as possible, I constrained myself to two-way models. I then discarded the runs that produced "tail probs" under 0.1 and retained less than 400K SNPs. Only a handful of models passed muster, including these two:

Armenian
Mycenaeans_&_Empuries2 0.233±0.041
Kura-Araxes_Kaps 0.767±0.041

chisq 18.422
tail prob 0.142151
Full output

Georgian
Globular_Amphora 0.071±0.025
Kura-Araxes_Kaps 0.929±0.025

chisq 18.419
tail prob 0.142266
Full output

At the most basic level, the results suggest that both Armenians and Georgians are overwhelmingly derived from populations of Bronze Age Transcaucasia associated with the Kura-Araxes archeological culture, albeit with minor ancestries from somewhat different sources from the west. As far as I can see, when using more than 400K SNPs and a wide range and large number of outgroups (or right pops), neither Armenians nor Georgians can pass perfectly for any one ancient population in my dataset.

The best proxies for the minor but significant western ancestry in Armenians are Mycenaeans of the Bronze Age Aegean region and Greek colonists from Iron Age Iberia (Empuries2). Obviously, and perhaps importantly, these are both attested Indo-European-speaking groups. On the other hand, the very minor western ancestry in Georgians is best characterized as gene flow from Middle to Late Neolithic European farmers rich in indigenous European forager ancestry. It's practically impossible to say what language or languages these farmers spoke. How about something Kartvelian?

In any case, for me, the perplexing thing about present-day Armenians is that they harbor very little steppe ancestry. By and large, no more than a few per cent. Compare that to the currently available samples from what is now Armenia dating to the Middle to Late Bronze Age, which show ratios of steppe ancestry of up to 25%. For now, I'm guessing that what we're dealing with here is the classic bounce back of older ancestry layers that has been documented for different parts and periods of prehistoric Europe.

See also...

Early chariot drivers of Transcaucasia came from...

Catacomb > Armenia_MLBA

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

286 comments:

1 – 200 of 286   Newer›   Newest»
Bob Floy said...

I thought that Georgians had almost no WHG?

Davidski said...

It seems like they do have a little bit.

Bob Floy said...

I'm confused about how Armenians could go from having 25% steppe ancestry to basically none, while retaining their(very unique) IE language, rebounding ancestry layers or not.

Slumbery said...

@Davidsky
"On the other hand, the very minor western ancestry in Georgians is best characterized as gene flow from Middle to Late Neolithic European farmers rich in indigenous European forager ancestry. It's practically impossible to say what language or languages these farmers spoke. How about something Kartvelian?"

There is nothing in these results that would suggest a Kartvelian or related language for those farmers. In your test Georgians are 90+ Kura-Araxes and as far as I know their uniparental markers are pretty compatible with Kura-Araxes as well. There is no reason to assume that the less than 10% European farmer ancestry (which we do not even know when and how arrived) caused a language shift. It is orders of magnitude more likely based on this that Kartvelian had its roots in Kura-Araxes.

Davidski said...

@Bob

Armenians actually seem to have retained a lot of steppe-derived paternal ancestry from the Middle to Late Bronze Age Armenian populations. Note the high frequencies of R1b-M269 in many parts of Armenia today.

But, in fact, the mainstream theory of how Armenians came to speak Armenian posits that there was a migration of Proto-Armenian speakers from the Balkans to the Caucasus during the Iron Age.

So my qpAdm results actually fit quite well with mainstream linguistics, because a pulse of ~20% admixture can be assumed to correlate with a language shift.

Davidski said...

@Slumbery

There is no reason to assume that the less than 10% European farmer ancestry (which we do not even know when and how arrived) caused a language shift. It is orders of magnitude more likely based on this that Kartvelian had its roots in Kura-Araxes.

Yes, I agree. I was just trying to stir up a debate about the origin of Kartvelian.

Cpk said...

It is possible that Kura-Araxes brought Chg and Hattian language. Maybe people there were already speaking Indo-European.

Davidski said...

@Cpk

The three Indo-European languages in the Caucasus - Armenian, Ossetian and Russian - are all late arrivals there.

Davidski said...

Just had a thought: the European farmer-like admix in Georgians might just be a reflection of a higher cut of Dzudzuana Ice Age ancestry?

Dzudzuana Ice Age foragers: a different type of Caucasus hunter-gatherer (Lazaridis et al. 2018 preprint)

Aram said...

Armenians as a Caucasian nation is a recent classification based on modern political boundaries.
Historical Armenia included western territory. Modern eastern Turkey. Which is off course contingent to Anatolia. I suspect one will easily get the same result if he use Anatolia_MLBA instead of Myceneans. So this is just an extra ANF/EEF that Armenian has relative to other Caucasians. Which partly is due to geography. Partly other factors.

The source of this extra ANF/EEF is complex. For example Armenians have a lot off J1-P58 ( some 7%) which is associated with Semitic people expansion and thus Levant_BA. Armenians have also other y dna associated with Levant_BA type ancestry. Like J2b1 and maybe some lineages of T. This has historical reasons. For example the politic of Artashecids to relocate Levantine people into Armenia.

If someone tries to take into account this Levant_BA type ancestry then the model will start to need something like Armenia_MLBA.

Samuel Andrews said...

I agree with Aram. Impossible to say if Armenians have extra Anatolian-ancestry due to an ancient cline running from Anatolia to the Caucasus or have recent southern European ancestry.

Bob Floy said...

@David

"the mainstream theory of how Armenians came to speak Armenian posits that there was a migration of Proto-Armenian speakers from the Balkans to the Caucasus"

IIRC it's supposed to be connected to Phrygian, also probably from the Balkans.

Aram said...

Diakonoff who proposed the migrationary theory of Armenians linked proto-Armenians with Mushki and Urumu tribes attested at LBA to Early Iron Age transition period circa 1176 BC. Back then there was no much information about archaeological situation in that region of Western Armenia and Cappadocia that were the main places of activities of this tribes.
In recent decades a lot off effort was made to understand the situation in Early Iron Age in that region and we have very interesting results.
After the collapse of Hittite empire a new type of pottery appears in it's eastern parts of plain Anatolia ( do not confuse with eastern mountainous Anatolia) and Cappadocia. This pottery is called groovy or grooved ware. Not only a new type of pottery appears but also a different lifestyle. New settlements. Increase of demography.
This indicates that new people appeared in these regions and it was just a matter of time that someone will link this new changes to Mushkians.
By the way this was done by a Turkish scholar. So no reason to think that he has a bias.
https://www.jstor.org/stable/3642931?seq=1#page_scan_tab_contents

According Veli Sevin Mushkians (grooved ware) came from Caucasian side and ultimately from Steppe.
Other scholars like Bartl also favoured an eastern origin. Others proposed a local origin of this pottery. The most important thing is that _NO_ one proposed a Western Anatolian origin of this pottery. And as far as I know no one denies the link between Mushkians and grooved ware. Only more nuanced proposals were made like that groovy ware can be a multiethnic phenomena.

This is corroborated by the genetic data. One can easily notice that those Anatolia_MLBA samples have lower CHG shift. Which means extra migrations from east occurred after the Late Bronze Age. At Iron Age for example.

Davidski said...

@Aram

I suspect one will easily get the same result if he use Anatolia_MLBA instead of Myceneans.

Nope, the model fails badly.

Aram said...

Davidski

So You propose to believe me that Armenians do not have ancestry from Anatolia_MLBA? Seriously?

Bob Floy

There a new theories out there. Scholars who follow this topic in general favour Armeno-Thracian hypothesis and a link with Phrygian

--------

For a long time a Thraco-Phrygian hypothesis grouping Thracian with the extinct Phrygian language was considered, largely based on Greek historians like Strabo. By extension of identifying Phrygians with Proto-Armenians, a Thraco-Phrygian branch of Indo-European was postulated with Thracian, Phrygian and Armenian and constituent languages. The evidence for this seems to have been mostly based on interpretations of history and identifying the eastern Mushki with Armenians and assuming they had branched off from western Mushki (whom has been conclusively identified as Phrygians).[35] However, in 1988 Fredrik Kortlandt argued, on linguistic grounds such as a common treatment of Proto-Indo-European glottal stops, that Armenian descended from a Thracian dialect. Thus, forming a Thraco-Armenian branch of Indo-European. In 2016 Kortlandt extended his theories, postulating a link between Thraco-Armenian and the hypothetical Graeco-Phrygian language family, despite Thracian and Armenian being Satem languages and Greek and Phrygian Centum languages Kortlandt identifies sound correspondences and grammatical similarities postulating a relationship between his Thraco-Armenian family and the more established Graeco-Phrygian family. Graeco-Armenian is by itself a common hypothesized subgrouping of Indo-European languages. Kortlandt considers Albanian a descendent of Dacian which he regards as belonging to a separate language family as Thraco-Armenian.[36]

Older textbooks grouped Phrygian and Armenian with Thracian, but the belief is no longer popular and is mostly discarded.[37] Today, Phrygian is not widely seen as linked to Thracian.

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

Aram said...

Ah this part

..On the other hand, the very minor western ancestry in Georgians is best characterized as gene flow from Middle to Late Neolithic European farmers rich in indigenous European forager ancestry. It's practically impossible to say what language or languages these farmers spoke. How about something Kartvelian?..

If Kartvelians are from GAC then Basques can be from Funnelbeaker.

Bob Floy said...

@Aram

The experts seem to be forever confused about that issue, since the evidence is so scant when it comes to paleo-Balkan languages.

Dragos said...

Aram
Phrygian seems very close to Greek. They’re far more likely to be related; and possibly proto-Armenian too
Thracian is poorly attested. Difficult to say much about it

Bob Floy said...

@David

"Just had a thought: the European farmer-like admix in Georgians might just be a reflection of a higher cut of Dzudzuana Ice Age ancestry?"

That would make Georgians one of the most indigenous populations in Europe, interesting thought.

Davidski said...

@Aram

So You propose to believe me that Armenians do not have ancestry from Anatolia_MLBA? Seriously?

Of course they have it, probably in varying amounts, but it's not enough to have an impact on the models that I tried with these Armenian sequences from the EGDP and SDGP. And even if it's more than 20%, then it still only creates the same effect as Kura-Araxes_Kaps + Mycenaeans.

And I did try adding ancient samples from the Levant into three-way models, but this didn't create any statistically sound models with Armenia_MLBA.

Honestly, based on this experiment, I wouldn't be surprised if in most of Armenia all you have is paternal ancestry from MLBA populations, and almost all of the genome-wide ancestry is from the local Neolithic and post-BA waves from the west.

Aram said...

Dragos

Well it is exactly Thracian type names that are attested in ancient Armenia.
A Thracian name Diuzenes and a name Diucini attested in Urartian texts.
Diucini and Diuzines are both the same as Greek Diogenes but with one difference they had a specific shift g->c ( specific to Armenian ) and a similar shift to fricative in Thracian.
So Kortland's theory is more realistic from linguistic point of view that anything proposed until now.

Davidski

My ancestors are from West. But they moved to modern Armenia in 19th century. And 75% of modern Armenia comes from western Armenia because of politics of persecution and killings in Ottoman Empire.

And if Your modeles can't show any ancestry from Levant this means that this type of exercises are not reliable because they contradict the available uniparental data. And historic data also.

For example the migration of Armenians to Cilicia in middle ages and their partial back migration. Also attested presence of Syrians and Jews in Armenia. Why You think there is no such a thing that Armenian_Jews. Because they were assimilated.


It is also possible that we have no good samples from Levant. For example samples from North East Levant can be much more informative. That is why ancient DNA is needed from periods when we know for sure that Armenian was spoken. And as far I know such a research is in process and it will yield very interesting results. It is just a matter of time that opinion in West starts changes.

Davidski said...

@Aram

For whatever reasons, Kura_Araxes_Kaps and the Mycenaeans and Greeks totally cover the minor Levant ancestry in these Armenians, if they have it.

For instance, keep in mind that the Iberian Greeks are actually from a colony from western Anatolia.

Ric Hern said...

R1b-M269 equals Indo-European ? Never ! Heheheeh...

Aram said...

Davidski

///For instance, keep in mind that the Iberian Greeks are actually from a colony from western Anatolia.///

And this basically confirms what I was saying. That this is a Anatolian ancestry rather than anything from Balkanes. Which is further corroborated by the fact that both those ancient Greeks from Spain had a haplogroup J of Anatolian origin.

And Kura_Araxes_Kaps do not represent the whole Kura_Araxian ancestry from historic Armenia. Let's wait Kura_araxian period samples from Arslan-tepe and we will have more accurate models.

Davidski said...

@Aram

And this basically confirms what I was saying. That this is a Anatolian ancestry rather than anything from Balkanes.

Nope, these Greeks are clearly mostly of Mycenaean origin.

But it doesn't matter that they're in large part of western Anatolian origin too, because genes don't speak languages, people do, so all we really need to corroborate the theory that Armenian came from the Balkans is evidence of a population movement from the direction of the Balkans into Armenia.

Dragos said...

@ Aram

Arguably, there seem to be a Balkan signal in Armenians

Armenian
Armenia_EBA 51.75%
Levant_BA_North 29.8%
Balkans_ChL 9.3%
Gonur1_BA 4.7%
Armenia_MLBA 4.45%
Anatolia_EBA 0%
Sintashta_MLBA 0%
Catacomb 0%

Distance 1.3646%


Armenian
Armenia_EBA 51.15%
Levant_BA_North 30.6%
Balkans_IA 12.95%
Gonur1_BA 4.95%
Balkans_ChL 0.2%
Anatolia_EBA 0.15%
Yamnaya_Samara 0%
Armenia_MLBA 0%
Catacomb 0%

Distance 1.3173%

Interesting that Armenia MLBA is so peripheral . Probably need some Iron Age Armenia genomes to be clear

Matt said...

Models using Armenia_MLBA rather than KA as an intermediate step make most sense, yeah. I think it can be assumed that the shift to Armenia_MLBA captures language transition.

The Fsts for present day Armenian vs Armenia_MLBA show displacement towards the Levant, which makes good sense historically, as post Armenia_MLBA probably sees further moves towards more general merging of the West Asian pool.

Davidski said...

@Matt

Models using Armenia_MLBA rather than KA as an intermediate step make most sense, yeah. I think it can be assumed that the shift to Armenia_MLBA captures language transition.

It doesn't work though, even if it makes most sense.

I don't think Armenia_MLBA contributed much to the modern Armenian gene pool.

Aniasi said...

I think I have an idea on why Armenian shows such little steppe ancestry today, but had a higher percentage in the past.

Today's Armenia is actually the far periphery of the historic Armenia. When one looks at the Armenian population throughout history, it was historically centred toward Lake Van.

Due to changes and events over the past few centuries, especially those of the early 20th century, today's Armenia is a small northeaster portion of a much large historic region.

If modern samples come from from that population, it is no surprise that they would show much less steppe ancestry, since they are from neighbour-joining areas with less heartland ancestry.

Dragos said...

If we exclude Armenia EBA and do another run, Anatolia EBA seems to counter the loss, fit worsens somewhat.

Armenian
Anatolia_EBA 35.3%
Armenia_MLBA 28.7%
Levant_BA_North 23.1%
Gonur1_BA 12.9%

Distance 2.1013%

Matt said...

Hmmm.... I think you are correct that one way or another Armenia_MLBA is pretty distant, with no plausibly equally distant other West Asian relatives and is so hard to fix as a major source and is therefore unlikely to make a large contribution.

Visualised through G25 reprocessing - https://imgur.com/a/3Hl8Sry

However, in terms of timing, note that Armenia_IA_Lchasen_Metsamor plots pretty close to Armenia_MLBA, with a dose of shift towards Iranians. This is far from a transect but should inform ideas of shifts over time.

a said...

Interesting conclusions in some of your posters comments. They seem to contradict [obviously not well read in genetic fields.]earlier genetic findings.

Synome said...

These models support the Greco-Armenian hypothesis, which is also supported by the Chang tree. Nice to see.

Leron said...

All in all you can’t take the results from these tools as absolute. They can be hit or miss, although usually better at general indications than at detailed descriptions.

There’s unique differences between modern Republic of Armenia/Hayastan, classical Greater Armenia (encompassing a majority of Transcaucasia), and the inner/outer Kura-Araxes zone that used to cover the same region.

Inner KA are obviously related to Neolithic north Iran populations. Their descendants, Bronze Age Hurro-Urartian groups, occupied the southern portion of Transcaucasia. They were very distinct from the ancestors of Kartvelians who were more autochthonous to the region. There was a clear separation in language, religion, culture, and YDNA between them. Outer KA were more mixed, in particular with steppe and Anatolian. The Nakh peoples can be considered a living representative of them and modern Armenia a frontier between inner/outer KA. However, most of the outer KA seem to have been overtaken by the Iron Age Urartians who expanded into Kartvelian territory.

By the later part of the Iron Age the inner KA zone slowly started to collapse. Semitic groups began to occupy what would later become the southern part of classical Greater Armenia. From the north and east, Scythian and Medo-Persians infiltrated, and from the west the Balkanic related IE groups. Ever since the region became heavily mixed and fragmented to the point that a Transcaucasian heritage couldn’t be distinguished except in the northern periphery.

Modern Armenia is not a good representative of the old inner KA, but better as a western (Luwic/Balkanic) + outer KA descendant, in addition to a sprinkle of Turkic, Syrian, and Iranian elements.

Slumbery said...

@Davidski

How many models have you tried for Georgians? In G25 nMontes I can't do the fine filtering you did there, but I tried out a few things. In those tests Globular Amphora does not stand out as a source better than anything else. Also if Anatolia_MLBA is included, European sources pretty much go down to noise level without any fit problems. (Georgian_Laz and Georgian_Imer behaves differently tough. Laz has much more Anatolian and some more European.)

"sample": "Georgian_Laz:Average",
"fit": 2.2957,
"Kura-Araxes_Kaps": 72.5,
"Anatolia_MLBA": 25,
"Poland_GAC": 2.5

"sample": "Georgian_Imer:Average",
"fit": 2.8868,
"Kura-Araxes_Kaps": 92.5,
"Anatolia_MLBA": 7.5,
"Poland_GAC": 0

Slumbery said...

@Dragos

In Global25 nMonte with similar reference populations the tests do not pick up Balkan ancestry.

A simple one:
"sample": "Armenian:Average",
"fit": 2.3507,
"Kura-Araxes_Kaps": 52.5,
"Anatolia_MLBA": 47.5,
"Balkans_ChL": 0,

And a more crowded one:
"sample": "Armenian:Average",
"fit": 1.9315,
"Armenia_MLBA": 29.17,
"Levant_BA_North": 25,
"Armenia_ChL": 21.67,
"Anatolia_MLBA": 17.5,
"Anatolia_EBA_Ovaoren": 6.67,
"Balkans_BA": 0

And fit-wise including Iran is the winning move:
"sample": "Armenian:Average",
"fit": 1.0608,
"Hajji_Firuz_ChL": 43.33,
"Armenia_ChL": 15.83,
"Anatolia_MLBA": 15,
"Armenia_MLBA": 14.17,
"Levant_BA_North": 6.67,
"Anatolia_EBA_Ovaoren": 5,

Davidski said...

@Slumbery

I get basically the same outcome as before with qpAdm when I add Anatolia_MLBA.

https://drive.google.com/open?id=1fXbv4tfaajeg9RCznkY2ikq2JQuAmmtp

See that's the thing with the Global25; it can be used for much more complex models, but because of the way it works, it's more susceptible to overfitting, without any signals from the software that there's a problem, so you gotta be careful.

Slumbery said...

Also it might worth to step back to earlier references and see how modern Armenians react to Iran a source vs. how Armenia_MLBA reacts.

"sample": "Armenia_MLBA:Average",
"fit": 1.55,
"Kura-Araxes_Kaps": 43.33,
"Catacomb": 23.33,
"Armenia_ChL": 15,
"Hajji_Firuz_ChL": 14.17,
"Levant_ChL": 2.5,
"Barcin_N": 1.67,

"sample": "Armenian:Average",
"fit": 1.1769,
"Hajji_Firuz_ChL": 42.5,
"Kura-Araxes_Kaps": 23.33,
"Armenia_ChL": 15.83,
"Levant_ChL": 10.83,
"Catacomb": 4.17,
"Barcin_N": 3.33,

So the local ancestry that bounced back was probably not so local by a strict definition of the term.

HAUMAVARGĀ said...

@Davidski what about Kurds?

Arza said...

@ Davidski

Can you try this model for Georgians (Imer)?

Kura-Araxes_Kaps
Minoan_Lasithi
Russian_Smolensk

Arza said...

@ Slumbery

Crucial question - nMonte3 with penalty?

AWood said...

Greeks and Armenians both carry similar branches of R1b-Z2103 and J2 to the best of my knowledge. So regardless of the steppe signal in BA Armenia, we don't know whether all the R1b arrived at that time, or at a later time from the SE Balkans. Same could also be said for the J2, despite the fact J2 and R1b were found in ancient samples of Armenians.

Slumbery said...

@Daviski

Fine, including Anatolia_MLBA brings the dark shadow of over-fitting upon us, because of messy cross-connections with Kura-Araxes. However my other point was that I can't see GAC being better than other European sources.

For example I can swap it to Tisza_LN or Trypilai and nothing really changes. (I used Gerigian_Laz because it has much more non-KA ancestry than Imer.)

"sample": "Georgian_Laz:Average",
"fit": 2.581,
"Kura-Araxes_Kaps": 88.33,
"Poland_GAC": 11.67,

"sample": "Georgian_Laz:Average",
"fit": 2.3528,
"Kura-Araxes_Kaps": 86.67,
"Tisza_LN": 13.33,

"sample": "Georgian_Laz:Average",
"fit": 2.4025,
"Kura-Araxes_Kaps": 85.83,
"Trypillia": 14.17,

So it is not very specific and could easily came from many sources.



@Arza

Yes. I am using the online nMonte runner with the default 0.001 penalty.

Davidski said...

@HAUMAVARGĀ

I'll do a post about Kurds at some point soon.

@Arza

Kura-Araxes_Kaps 0.890
Minoan_Lasithi 0.044
Russian_West 0.066
chisq 20.281
tail prob0.0619561

Arza said...

@ Slumbery

Turn it off.

The whole concept of applying penalty according to distance is dubious to say the least. Additionally it's wrongly implemented and it randomizes results to a certain degree.


"sample": "Georgian_Imer:Average",
"fit": 2.8868,
"Kura-Araxes_Kaps": 92.5,
"Anatolia_MLBA": 7.5,
"Poland_GAC": 0

vs.

Georgian_Imer
Kura-Araxes_Kaps:ARM002-003 49.7%
Kura-Araxes_Kaps:ARM001 45.4%
Poland_GAC:N38 4.5%
Poland_GAC:I2403 0.4%
Poland_GAC:I2433 0%
Poland_GAC:I2441 0%
Anatolia_MLBA:MA2200 0%
Anatolia_MLBA:MA2203 0%
Anatolia_MLBA:MA2205 0%
Anatolia_MLBA:MA2206 0%

Distance 2.7381%

Andrzejewski said...

@Leron So you are saying that the Hurro-Urartians are descendants of the KA. But then you allude to the Nakh people also being their descendants. Am I correct?

If so, then the relatedness or the Hurro-Urartians to modern Chechen or North Caucasus (the Aloradian theory) was already rejected by most linguists.

And if you posit the KA is an offshoot of Iran_N, do you think that HU might be related to Elamite, Dravidian and/or BMAC language(s)?

MOCKBA said...

Gamkrelidze makes a big point about PIE Urheimat by documenting supposed borrowing / influences from Proto-Kartvelian to PIE. But the issue may be that with few, and not too deeply divergent, branches of Kartvelian languages, the proto-Kartvelian may only be reconstructed to comparably recent times vs. the genesis of PIE.

Andrzejewski said...

@Davidski "On the other hand, the very minor western ancestry in Georgians is best characterized as gene flow from Middle to Late Neolithic European farmers rich in indigenous European forager ancestry. It's practically impossible to say what language or languages these farmers spoke. How about something Kartvelian?"

So Kartvelian language family arose out these possibilities:

1. A completely local development, i.e. Dzudzuana/Darkveti Meshoko/Paleolithic Caucasus language.

2. Some Anatolian Farmer language coming either directly from Turkey or more likely from Cucuteni Tripolye or GAC.

3. As the EEF was WHG-rich, Kartvelian might be a WHG language long extinct in Europe proper.

Which one hypothesis seems more plausible to you?

Andrzejewski said...

@MOCKBA If anything, most borrowing was PIE -> Kartvelian, e.g. "heart" KRD -> MKRD.

Andrzejewski said...

@All Will Georgians be considered to be mostly CHG-like population vis-a-vis ANF/EEF?

How would KA fare? According to @Leron it is a Northern Iran-derived population. Would you classify Kura-Araxes as Iran_N rather?

Last time I checked KA/HU were approximately 53% ANF aDNA...

Last but not least - how does the recent finding fare regarding the Uruk expansion theory? Would Georgians/Kartvelians be Sumerians/Ubaidian immigrants to work at metallurgy, or is the Uruk Expansion theory completely discarded now? I am also curious how Semitic languages came about, considering that Semitic speakers have had the uniparental marker J1/J2 since at least the BA, largely replacing the Natufian/Iberomasulean E1b1 (which is very common in Greece and the Balkan BTW, and arrived with Neolithic farmers THERE), and also there was a study by Iosif Lazaridis BROAD recently that found out that there was a huge population replacement event taking place in Syria and Israel 6.5ybp, in which an Anatolian Farmer-like population possessing blue eyes and Neolithic Haplogroup T essentially replaced E1b1 male lines in the autochthonous population, only to vanish off the face of the earth 2 millenia later with a CHG like uniparental markers as typical Semitic speakers...

Slumbery said...

@Arsa + Davidski

Indeed, turning off the nMonte penalty increases GAC ancestry in the presence of Anatolian_MLBA.

"sample": "Georgian_Imer:Average",
"fit": 2.7378,
"Kura-Araxes_Kaps": 95,
"Poland_GAC": 5,
"Anatolia_MLBA": 0,

"sample": "Georgian_Laz:Average",
"fit": 2.1941,
"Kura-Araxes_Kaps": 75.83,
"Anatolia_MLBA": 17.5,
"Poland_GAC": 6.67,


It does not change the lack of the specific reaction to GAC however. A lot of European farmer groups fit just as well or better. And Georgian_Laz has some more Southern ancestry (I do not not think it is 15-20%, but it is there).

Slumbery said...

@Andrzejewski

You should forget seeking roots of modern languages in the time-depth of Dzudzuana. It is pointless. It is like three times of the age of the oldest extant language families. As far as language is concerned it is completely beyond even a speculative reach.

BTW Dzudzuana is more like a cousin of the ancestors of Anatolians than simply an ancestor to CHG. ANF is closer to it than CHG is, because CHG had a lot of ANE mixture.

And Kura-Araxes is a multi-way mixture in Neolithic time-depth. It is CHG + Anatolian + Iran + probably even some Levant. The entire region in a turmix. I'd say probably 1/3 ANF (give or take).

Arza said...

@ Davidski

Thanks.
Russian admixture wasn't rejected, but it seems that it didn't play a major role. Maybe some population from between Barcin and LBK would work better than Minoans in this setup. If not, my vote for a WHG source goes to Dzudzuana/PinarBasi.

That's how AHG looks in the original PCA from the paper:

PinarBasi
Anatolia_N:I1097 50.0%
Anatolia_N:I1581 38.2%
Iron_Gates_HG:I4873 11.2%
Anatolia_N:I0707 0.6%
Distance 0.9168%

For example distance between Anatolia_N:I1581 and Anatolia_N:I1583 in this PCA is 1.7098%.

Slumbery said...

@Arza

"If not, my vote for a WHG source goes to Dzudzuana/PinarBasi."

I find this quote unlikely. Again, too much time. There is no way a ghost population with meaningfully specific roots in Dzudzuana survived into Kura-Araxes times in that region. In a lesser extent the same is true for PinarBasi. How would it turn up in Georgia 11 000+ years later, while going under the radar in the entire region until that?

There are a lot of possible routes for this minor WHG got there. Bronze and Iron Age + antique connection to the Balkan for example.

Leron said...

@Andrzejewski:

From historial, linguistic and archaeological perspectives, Hurro-Urartian people were closest to proto-Kassites (while they were in the western Zagros, prior to entering Babylon) and the Manneans of northwest Iran.

However, the early ancestors of Nakh people, including the Chechens, were a seperate group (J1 males) that mixed with incoming Kura-Araxes groups from the east (the north Iran area and hence their link to BMAC) and some steppe from the north. I believe there is a linguistic connection between Northeast Caucasian (Nakh) to Hurro-Urartian and related languages that was probably due to the cultural dominance of Kura-Araxes. The Nakh being more receipt than a descendant of the originators.

Elamites and Dravidians are separate entities in respect to the Kura-Araxes people. Geographically they connect with the southern Iranian lands and Balochistan than anywhere in north Iran.

Arza said...

@ Slumbery
We need to wait for genotyped data and G25 coordinates, but from the 10-dimensional PCA published in the original paper it looks like PinarBasi is virtually identical to LBK_N, so there is a possibility that EEF were already WHG-shifted when they were entering Europe.

It's possible that with more samples we will see more WHG-shifted (compared to Barcin) Anatolian Farmers in actual Anatolia.

Dragos said...

There might be at least a couple of issues with the Pinarbasi paper, as ssome of the conclusions are questionable at a basic level - such as the degree of continuity between AHG & Neolithic

At some point, there must have been some minor movement of WHG-like people from Europe to Anatolia - manifest by Y-hg I2c, but overall the degree of movement from the east is probably significantly under-estimated by there, because of lack of samples from Northern Syria, for ex.

Samuel Andrews said...

I added a MidEast slide to my WestEurasia G25 index. It models Mid East according to Caucasus_Eneo, Iran_Eneo, Levant_Neo, BarcinNeo-excess, Yamnaya, tons of other minor stuff.

https://docs.google.com/spreadsheets/d/1LPWAEC3dbAEDu8aBAAcxIOa5CQjuflt0f0cvhCpZ_ME/edit#gid=1497568895

Armenian & Georgian are pretty different in this model...

Georgian_Imer.
Caucasus-73%
Iran-7%
LevantN-5%
Barcinexcess-14%
Yamnaya-2%

Georgian_Laz.
Caucasus-61%
Iran-8%
LevantN-8%
Barcinexcess-20%
Yamnaya-3%

Armenian
Caucasus-26%
Iran-27%
LevantN-19%
Barcinexcess-21%
Yamnaya-6.1%

Geogrian_Imer is identical to Maykop. Armenian, is mostly Assyrian-like

Arza said...

The question about dating of burials from the upcoming megalithic paper suddenly became even more interesting as R.Rocca reports that:

Primrose9BE has a lot of DNA, but not a lot of Y-DNA, so while he is positive for these R1b SNPs, the sample may be female:

R1b1a-CTS3794
R1b1a1a-PF6501
R1b1a1a2-L777
R1b1a1a2-CTS329

PF said...

Looking at these two-way models, they're kind of counterintuitive because Kura-Araxes didn't reach a large part of modern Georgia, while being smack in the middle of modern Armenia. Also, the Kura-Araxes border maps quite well to where G2a increases heavily north of it (~1/3rd of modern Georgians, even higher in the west / north) while inside the borders there's a lot more J2.

I wouldn't say that Georgians and Armenians are *that* alike either. What seems to generally separate them is that the latter have a lot more Iranian and perhaps Anatolia_Chl related ancestry.

[1] "distance%=2.4868"
Armenian
Anatolia_ChL,51
Seh_Gabi_ChL,36.4
Darkveti-Meshoko,9.6
Ukraine_GAC,3

[1] "distance%=2.5859"
Georgian_Imer
Darkveti-Meshoko,71.2
Anatolia_ChL,15.8
Seh_Gabi_ChL,9
Ukraine_GAC,4

PF said...

Hmm, these models are more interesting. Good fits and kind of make sense?

[1] "distance%=1.7709"
Armenian
Kura-Araxes_Kaps,63.2
Levant_ChL,29.8
Vonyuchka_Eneolithic,7

[1] "distance%=2.0605"
Georgian_Imer
Maykop_Novosvobodnaya,83.6
Anatolia_ChL,16.4

Davidski said...

@PF

Looking at these two-way models, they're kind of counterintuitive because Kura-Araxes didn't reach a large part of modern Georgia, while being smack in the middle of modern Armenia.

Kaps is in northern Armenia near the Georgian border. This is also the area where some of the earliest Kura-Araxes sites are located, including across the border in Georgia.

Most of the Georgian gene pool might derive from this area.

Grey said...

Bob Floy said...
"I'm confused about how Armenians could go from having 25% steppe ancestry to basically none, while retaining their(very unique) IE language,"

intrusive military elite concentrated in strategic locations getting into a long and eventually losing war with a more powerful neighbor so the bulk of the steppe ancestry ended up massacred in their fortresses and towns while the only slightly admixed peasantry in the surrounding villages survived.

Hugh Capet said...

@Davidski

A bit unrelated, but I had an inquiry. I think these results are fairly accurate when it comes to formal statistics. The question I had in mind was, what differentiates Armenians and Georgians from Iranians, and why are the latter considered a Near Eastern population as opposed to a Caucasian population (from the Caucasus). Iranians are mostly descended from the Neolithic era inhabitants of Iran, and thus have a similar ancestral make-up to Armenians and Georgians, both of whom are mostly ancestrally derived from Caucasus Hunter Gatherers. Furthermore, Semitic related admixture in Iranians is no more than 5% on average. Thus, would not it be more accurate to classify Iranians as a Caucasian population and not a Near Eastern one? I have noticed that on your Principle Component Analyses, Iranians are labelled as Near Eastern. Is this a matter of geography or is it due to the Neolithic population of Iran being considered Near Eastern itself? Regardless, I think it is safe to conclude that Iranians are a fairly distinct population from their Semitic-speaking neighbours. Thoughts?

Samuel Andrews said...

Kura-Axes is generally similar to but not ancestral to Georgians or Armenians. Geogrians are basically identical to Maykop.

Armenians are very similar to Assyrians & eastern Turkey (if you remove Turkic admix). So, they look more like a northern Mesoptamia/Eastern Turkey pop than a Caucasus pop.

Davidski said...

@Hugh Capet

Iranians are called a Near Eastern population in my analyses based on their geography.

I don't think it's reasonable to classify Iranians as a Caucasus population based on any criteria. That would be a highly original thing to do, and I generally don't like to stray too far from the mainstream on this blog.

If the labels in my analyses were based purely on genetics, and I was looking to differentiate Iranians from their non-Indo-European western neighbors, who do show somewhat different genetic structure, I'd probably call them a South Caspian population.

Having said that, there's no strict genetic border between the populations of Iran and other parts of the Near East.

They form a genetic continuum, or cline, from west to east based on somewhat different ratios of the same main ancient ancestral components, which are usually termed Anatolian-, Levant- and Iranian-related, respectively.

We know from ancient DNA that this continuum has existed in the Near East since at least the Epipaleolithic, which means that the people living within and outside the western borders of what is now Iran have been mixing with each other for thousands of years, even before the appearance of any of the linguistic groups that dominate the region today, such as Semitic and Indo-European.

Andrzejewski said...

@Samuel Andrews "Kura-Axes is generally similar to but not ancestral to Georgians or Armenians. Geogrians are basically identical to Maykop."

OK. What's the genetic composition of Maykop and how different is it from KA?

Does this new study support or debunk an "Uruk Expansion" theory?

(I read that Darkveti Meshoko and Shulaveri are very similar to Halafian and Ubaindian material culture).


Slumbery said...

@Samuel Andrews
"Kura-Axes is generally similar to but not ancestral to Georgians or Armenians. Geogrians are basically identical to Maykop."

It is more complicated than this. Georgians seem to have ancestry that is more on the Kura-Araxes side than on the Maykop side.

"sample": "Georgian_Imer:Average",
"fit": 1.8978,
"Maykop_Novosvobodnaya": 56.67,
"Kura-Araxes_Kaps": 36.67,
"Trypillia": 6.67,

"sample": "Georgian_Laz:Average",
"fit": 2.1389,
"Kura-Araxes_Kaps": 58.33,
"Maykop_Novosvobodnaya": 27.5,
"Trypillia": 14.17,


"Armenians are very similar to Assyrians & eastern Turkey (if you remove Turkic admix). So, they look more like a northern Mesoptamia/Eastern Turkey pop than a Caucasus pop."

Aren't they a bit low on Levant ancestry for Mezopotamia? I'd say they have a lot of ancestry from somewhere just a bit South, around the modern Turkey-Iran border region or so. The biggest excess ancestry they have over Armenia MLBA is seems to be Iranian.

Davidski said...

@Samuel Andrews

Kura-Axes is generally similar to but not ancestral to Georgians or Armenians. Georgians are basically identical to Maykop.

I'm pretty sure that both Armenians and Georgians are in large part of Kura-Araxes origin. That's probably one of the reasons why they like their wine so much.

https://www.youtube.com/watch?v=sWVY2g0JyjI

And the reason that Georgians are so similar to Maykop people is because the Maykop culture was mostly founded by migrants from Transcaucasia, and Georgians are the best living example of prehistoric Transcaucasians.

Andrzejewski said...

@Davidski "We know from ancient DNA that this continuum has existed in the Near East since at least the Epipaleolithic, which means that the people living within and outside the western borders of what is now Iran have been mixing with each other for thousands of years, even before the appearance of any of the linguistic groups that dominate the region today, such as Semitic and Indo-European."

Is that the reason why Levant_N and Anatolia_N are distinct but close populations on the cline? If I recall correctly about 1/3 of Levant_Epipaleolithic is Anatolia HG-derived, whereas after 7000 BC roughly 20% of Anatolia aDNA is Levant-based.

If so, does Levant dna within Anatolian farmer pops explain the abundance of E1b1b in Greek and Balkanic populations?

Bob Floy said...

@Sam

My gut is that proto-Kartvelian was spoken by Maykop.

Andrzejewski said...

@Sam Andrews @Bob Floy I thought that Maykop was ancestral to Northwest Caucasus language family rather than to proto-Kartvelian. Besides, some similarities between PIE and NW Caucasus language family may be attributed to the mutual influence exerted between Maykop and Steppe.

Are you hinting then that NWC and Proto-Kartvelian form a proto-phylum?

Davidski said...

@Andrzejewski

You're not allowed to use ethnic slurs in the comments here, even if you want to talk in "slang". It's against the rules.

Andrzejewski said...

@Davidski I will rephrase it then without the slang. It is known that Russians and Ukrainians tend to regard Armenians and Georgians as "different", and I am kind of wondering if the reason that the latter ones look different than Eastern Slavs is due to a strong cline of WHG-EHG (especially the EHG) that is no lacking in Caucasus people. I personally tend to think that the reason Europeans are so different from both West Asian and Caucasus people is because of the overwhelmingly dominant genetics of the Steppe Kurgan migrants, even in Southern Europe.

Bob Floy said...

@Andrzejewski

There are a lot of PIE/Kartvelian similarities also, at least according to some research.

Andrzejewski said...

@Bob Floy "There are a lot of PIE/Kartvelian similarities also, at least according to some research."

Would you agree with me that mostly it's due to a PIE -> Kartvelian borrowing?

Slumbery said...

BTW, adding Iran ChL as reference unveil a significant difference between Laz and Imer. The Laz samples of G25 are from Goergia, but the Laz are much more a population from Turkey and apparently have the same ancestry the Armenians have (but less of it), while the core Georgian population does not have it. The excess Anatolian and Iranian ancestry of the Laz (I mean excess above what the Imer samples have) probably came from gradual mixing with other Eastern Anatolian groups, like Armenians.

"sample": "Georgian_Imer:Average",
"fit": 1.8978,
"Maykop_Novosvobodnaya": 56.67,
"Kura-Araxes_Kaps": 36.67,
"Trypillia": 6.67,
"Hajji_Firuz_ChL": 0

"sample": "Georgian_Laz:Average",
"fit": 1.9778,
"Maykop_Novosvobodnaya": 39.17,
"Kura-Araxes_Kaps": 31.67,
"Hajji_Firuz_ChL": 16.67,
"Trypillia": 12.5

Davidski said...

@Andrzejewski

I personally tend to think that the reason Europeans are so different from both West Asian and Caucasus people is because of the overwhelmingly dominant genetics of the Steppe Kurgan migrants, even in Southern Europe.

Well you're wrong.

Europeans look more or less alike relative to most Near Easterners because we're very closely related, not just in terms of sharing the same main ancient ancestral components in similar ratios, but also due to relatively recent homogenizing factors, such as the Migration period, and even genealogical links.

On a global scale, Europeans are a remarkably homogeneous continental population, and this has had various by-products, one being that Europeans are overall relatively distinct from Near Eastern populations.

Bob Floy said...

@Andrzejewski

"Would you agree with me that mostly it's due to a PIE -> Kartvelian borrowing?"

I'm not expert enough to say with any certainty, but according to what I've read there are grammatical similarities in addition to shared lexical items.

Andrzejewski said...

There is an interesting theory right here about the Northwest Caucasus-like substrate in North West IE languages, under this entry: https://en.wikipedia.org/wiki/Nordwestblock

"linguist Peter Schrijver speculates on the reminiscent lexical and typological features of the region from an unknown substrate whose linguistic influences may have influenced the historical development of the (Romance and Germanic) languages of the region. He assumes the pre-existence of pre-Indo-European languages linked to the archeological Linear Pottery culture and to a family of languages featuring complex verbs, of which the Northwest Caucasian languages might have been the sole survivors."

So according to him the LBK spoke a Northwest Caucasus-like language, which is basically then an Anatolian farmer one. Which means that Adyghe and other NW Caucasus languages may be related to the languages spoken by first farmers of Europe. What corroborates this hypothesis is that Herodotos used to assert that the first inhabitants of Crimea were the Taurs which were linked somehow with modern NW Caucasus speakers.

Andrzejewski said...

@Bob Floy "I'm not expert enough to say with any certainty, but according to what I've read there are grammatical similarities in addition to shared lexical items."

It's intriguing that PIE was influenced with agricultural vocabulary by West Asian farmers and pastoralists. For example the word for cow, "guowus" seems to be a borrowing from either Sumerian or from some other non-IE language.

Slumbery said...

Sorry for the monologue, I post just another G25 nMonte test on the topic of "Southern" ancestry in Laz vs. Georgians from Imer. Apparently this is not as simple as Armenian mixing into Laz, because Laz react much better to Levant ancestry than to Iranian.

"sample": "Georgian_Imer:Average",
"fit": 1.8973,
"Maykop_Novosvobodnaya": 57.5,
"Kura-Araxes_Kaps": 35,
"Trypillia": 6.67,
"Levant_BA_North": 0.83,
"Hajji_Firuz_ChL": 0

"sample": "Georgian_Laz:Average",
"fit": 1.827,
"Maykop_Novosvobodnaya": 51.67,
"Kura-Araxes_Kaps": 25,
"Levant_BA_North": 13.33,
"Trypillia": 10,
"Hajji_Firuz_ChL": 0

"sample": "Armenian:Average",
"fit": 1.2749,
"Kura-Araxes_Kaps": 40,
"Levant_BA_North": 27.5,
"Hajji_Firuz_ChL": 15.83,
"Maykop_Novosvobodnaya": 9.17,
"Trypillia": 7.5

Bob Floy said...

@Andrzejewski

Yeah, but is there really a northwest Caucus-like substrate in Celtic and Germanic languages? I don't think there's much support for that idea. For a long time it was said that Celtic languages had a Semetic substrate, I saw it presented as fact once in a BBC article, but that idea dosen't stand up to examination. We know that there's some kind of substrate in Germanic, but I've never heard that it was supposed to be from the Caucuses. If any IE languages had a Caucasian substrate, it would have been acquired during the PIE phase, so wouldn't all IE languages have a Caucasian substrate, rather than just two families?

I'm always a little skeptical of those sorts of claims, it seems like they're a dime a dozen.

Bob Floy said...

And I'm skeptical that farmer groups like LBK spoke Caucasian languages, dosen't ring true for me.

Davidski said...

Apparently the words for clover in Celtic, Germanic and Georgian all come from the same, presumably Kartvelian source.

Why? Maybe Maykop?

Bob Floy said...

@Davidski

It could easily have been from contact with Maykop, but, again, it's widely held that there's vocabulary overlap between IE and Kartvelian, which is thought to be from deep in prehistory at the proto stage.

Andrzejewski said...

@bob "It could easily have been from contact with Maykop, but, again, it's widely held that there's vocabulary overlap between IE and Kartvelian, which is thought to be from deep in prehistory at the proto stage."

From the CHG in Kurgan populations of the Steppe?

Dragos said...

So when was Kartvelian spoken north of the Caucasus ?

Davidski said...

During the Maykop period. Otherwise, it's hard to explain how the word for clover spread from the Caucasus to Northwestern Europe.

Ric Hern said...

My personal view is that Languages change as everybody knows and similar innovations can occur thousands of kilometres from each other independently. Even neighbouring languages can come to exhibit the same eventual similarities without any significant contact. Languages for me is like Chess. The ultimate goal is the same, trying to transmit an idea as fast and as clear as possible. Some times it leeds to similar salutions to a problem among different languages. This does not necessarily mean that there had to be contact. And similarities with only a few words can point to many things example, trade which does not specifically mean direct contact. If you can pull out a dictionary and point to 1000 similar words or cognates it is a totally different ballgame...

Davidski said...

I don't know, I'm just messing around, but the Maykop samples do share a lot of drift with Georgians, and the word for clover is apparently just one example of these sorts of links between Kartvelian and certain Indo-European branches.

Bob Floy said...

If the "clover" example isn't an isolated thing, than it could well be part of that apparent vocabulary overlap, it may not be from literal Yamnaya/Maykop contact. If I understand right, the lexical similarities between IE and Kartvelian go back to the time of their formation.

Samuel Andrews said...

Btw, Maykop samples aren't all the same. Some are like Geogrians, some similar to Kura Axes.

Davidski said...

Yeah, probably because Maykop people came from different parts of the south Caucasus and surrounds.

Samuel Andrews said...

1.848"

Georgian_Imer

Maykop_Novosvobodnaya1,71.2
Anatolia_ChL,13.9
Anatolia_MLBA,6.2
Kura-Araxes_Kaps,4.4
Kura-Araxes_Talin,4
Anatolia_BA,0.3
Kura-Araxes_Kalavan,0
Kura-Araxes_Velikent,0
Anatolia_EBA,0
Levant_ChL,0
Levant_BA_North,0

1.5941"

Georgian_Laz

Maykop_Novosvobodnaya1,46.7
Anatolia_MLBA,22.2
Kura-Araxes_Velikent,11.9
Anatolia_BA,11.6
Anatolia_ChL,4.6
Kura-Araxes_Talin,3

1.3645"

Armenian

Kura-Araxes_Kalavan,24.6
Anatolia_MLBA,19.6
Levant_BA_North,19
Kura-Araxes_Kaps,15.3
Kura-Araxes_Velikent,13.7
Levant_ChL,4.5
Anatolia_ChL,2.9
Maykop_Novosvobodnaya1,0.4

Slumbery said...

@Samuel Andrews

Kura-Araxes samples are not completely uniform either and their range intersects with the range of Maykop. These two populations have a lot of recent common ancestry.

Aram said...

Davidski

You can't use Armenia_MLBA because the true counterpart of Armenia_MLBA is missing from Your datasets. This counterpart was living in South Western part of historic Armenia and modern Iraqi Kurdistan which back then was Semitic. Assyrian empire. Also we don't know how Aramaic people where looking who infiltrated very north. We have no single sample from this people.

But there is another way to know how much of this Armenia_MLBA is left in modern Armenians.
Many people didn;t like Busby's work when he said that Italy witnessed a large influx during the early centuries of Current Era. Maybe because he made a wrong statement that this influx came from Arabs. But now ancient DNA will exactly confirm what Busby and Hellenthal were saying.

Now the same Busby et al. has found an important long lasting admixture event in Armenians. Spaning from LBA to Hellenistic era and even more.

This admixture event had 30-35% contribution to modern Armenian and was modelled as a mixture of something Levantine + Kumyk. Now this Kumyk represents the Armenia_MLBA. Event by fst the Kumyks are one of closest to Armenia_MLBA. While Levant represents something from NE Levant.

This mean that modern Armenians have something like 15-17% from Levant ( which is nicely corroborated by Y dna ) plus something like 15% from Armenia_MLBA. Which represents mostly the Z2103, I2c2 and almost certainly E-V13 also. Because the highest level of E-V13 is found in Urmia bassein.

The rest of Armenian ancestry is from that Kura-Araxes to Anatolia_MLBA/EBA cline. Mostly local G2/J2 dna.

There is nothing from Balkanes in any of this papers. Neither in Marc Habers paper. Not only there is no true signals from Balkanes in autosomal level. But there is no single haplotype from Balkanes that can fit the bill.


Dragos said...

@ Slumberry

I think it's hard here.
For ex, you have Anatolia MLBA, which arguable has some Balkan admixture (or some of them). And its difficult if we are sourcing using a mixture of Iron Age & Chalcolithic samples, which makes any conclusion difficult

So what about using a more steady approach - of Chalcolithic - EBA sources only
We might get:

Georgian_Imer
Maykop_Novosvobodnaya 93%
Barcin_N 7%
WHG 0%
Lokomotiv_N 0%
Armenia_EBA 0%

Distance 1.8368%


Armenian
Armenia_EBA 56.3%
Levant_BA_North 31.4%
Varna 7.8%
Maykop_Novosvobodnaya 3.6%


Distance 1.4068%


Nothing definitive, but it seem most agree that Georgians have a strong pull to Majkop, whislt Armenians do not; and the Balkan signal keeps appearing. IIRC Aram stated one of the said descendants of Armenian noble rulers is supposed to have been I2c


@ Davidski

''During the Maykop period. Otherwise, it's hard to explain how the word for clover spread from the Caucasus to Northwestern Europe. ''

Possibly (see above^), or maybe it's just a wonderworter.

Dragos said...

Davidski
Can you do IBD or haolotype analysed ?

Aram said...

This is the link for Busby

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4714572/

Dragos

There can be a signal from Balkanes on haplotype level. Specifically from North East Balkanes where the relatives of Armenians the Thracians were living. (Predominantly R1b-Z2103 and some I2 in that Scythian paper) And this "signal" was buried in Lchashen Metsamor burial in elite burial with wagons.
Needless to say that this Balkanic signal has nothing to do with the extra ANF that Armenians had relative to Georgians just due to geography.
I understand that You don't believe that Catacomb/Multicordoned ware are the source of Mycenean Greek language.
But You don't propose any alternative to Yamna. Are You going to suggest that Greek is a EEF language?

Davidski said...

@Aram

That Busby paper is based on modern data and an outdated algorithm that didn't get things exactly right.

Bob Floy said...


@Andrzejewski

"From the CHG in Kurgan populations of the Steppe? "

It could even be that IE and Kartvelian are genetically related, if we go back far enough. If people are now proposing that Basque could be related to IE, then this is certainly no more wilder than that. But my guess would be that the relatedness stems from contact between the two in their formative stages.

Dragos said...

@ Aram

''I understand that You don't believe that Catacomb/Multicordoned ware are the source of Mycenean Greek language. ''

Indeed, the Multi-Cordoned ware didnt even make it to Wallachia, let alone to GReece, so to beleive Mycenean elites came from there relies on ''belief'' ;)


'' Are You going to suggest that Greek is a EEF language? ''

I wouldn’t suggest something as basic as that. “EEF languages, steppe langages, CHG languages”. That's baby talk bro

Aram said...

Busby correctly predicted the migration to Italy from Near East in Imperial period.

And You also use modern Armenians for Your modeling. Why You do don't do the same to Iron Age Armenian who had a European haplotype I2c2 and was found in elite burial.

Davidski said...

Busby also incorrectly claimed that there was Mongolian ancestry all over Eastern Europe.

So the saying about that blind chicken comes to mind.

Aram said...

Busby claimed that in an inteeview to journalists. Not in their paper. As far as I remember.
They have found a Mongolian admix in some North Caucasians and maybe some groups who had a Turkish admix.

Slumbery said...

@Dragos

We agree on the much more Northern roots of Georgians. As for the Balkan signal in the Armenians: of course it appears when the only reference population in test that could convey similar ancestry is from the Balkan. The Balkan - Anatolia_MLBA interaction is a problem both way, it is hard to separate their possible effects on Armenians.

But I let go this line of discussion. I admit that I have no strong genetic argument against some Balkan genetic source into Armenians. The least we can say that the idea is not contradicted by the data.

Davidski said...

@Aram

Enough already. Those papers based on modern haplotypes are crap.

One of their papers showed a massive pulse of admixture from West Asia into Poland dating to the medieval period.

Just move on and use ancient data to make any point that you want to make.

Matt said...

@Aram, yeah, in hindsight it's odd how much controversy it feels like there was about Busby's West Eurasian admixture landscape paper, since it seems like it will be largely validated.

I still raise an eyebrow at their Levant->France and West Africa->Basque donor profile though. I'd love to have seen their reconstructed sources projected on something like the G25, might have made us go "Ah, that source is actually....".

Aram said...

Davidski

Ancient DNA already demonstrated that there was a migration to Armenia from East Europe during MLBA. The bulk of impact probably occured at LBA.

How much modern Armenians have steppe ancestry is absolutly unimportant.

@All

It will be good if Your post a model for Anatolian Cappadocian Greeks.

Gaska said...

@Arza said...The question about dating of burials from the upcoming megalithic paper suddenly became even more interesting as R.Rocca reports that: Primrose9BE has a lot of DNA, but not a lot of Y-DNA, so while he is positive for these R1b SNPs, the sample may be female:

R1b1a-CTS3794
R1b1a1a-PF6501
R1b1a1a2-L777
R1b1a1a2-CTS329

It may be due to contamination by handling samples in Spain there is a case of R1b-M269 in the Neolithic discarded for this reason. In another case it would be impossible for a woman with these SNPs, what explanation has Rocca given? I do not think it takes a long time to publish the paper and know the dates.

@Aram There is some paper published on the south of Italy in the imperial period (2nd century AD) that shows that Levantine influence. It is normal, the population movement in Roman times was very large. What happens is that they did not study male markers but Mit-groups.

Aram said...

It will be good idea to model modern Cappadocian Anatolian Greeks using this source population.


sample": "Armenian:Average",
"fit": 2.3507,
"Kura-Araxes_Kaps": 52.5,
"Anatolia_MLBA": 47.5,
"Balkans_ChL": 0

To make sense how modern Armenians are looking one need to k

Dragos said...

@ Slumberry

I don't have any clear idea about proto-Armenians either. It's hard to know much detail about the post-B.A. of that region. There'd be a lot to learn, most of it in non-English or European languages.
But Myceneans are easier to guess. Solving a lot of problems also is that Phyrgian, Macedonian & Greek are obviously related, and likely diverge at western Macedonia/ Epirus; whilst Thracian & Dacian form their own clade, with the latter being a younger offshoot, from the south.In the Iliad, it is written that the Phrygians and Mushki fought under a common general.

PF said...

While Maykop and KA are generally similar, I mentioned the known KA border because there does seem to be something going on related to it that's difficult to ignore. There's a reason why Georgians greatly prefer Maykop_Novosvobodnaya while Armenians don't need it at all...

One of the main relevant mysteries here is Kartvelian. I don't think it's an accident that where Kartvelian is spoken today, and more so, where Svan, the oldest branch, remains, corresponds so well to very high amounts of G2a. Not just where Kartvelian survives but in the NW Caucasus in general, irrespective of current language (Abkhazians, Ossetians, etc.).

Interestingly, I also noticed that in this area, where there is the highest amount of G2a, is also where there is the highest amount of CHG proper ancestry. In G25 the closest modern populations to CHG are Abkhasian, North_Ossetian, and Georgian_Imer.

What this is telling me is that this is reflecting the original neolithicization of the area, when G2a-rich early farmers mixed with straight up CHG, forming something like proto-Eneolithic_Caucasus. Since then the entire region was influenced my movements from all directions, including e.g. south from KA and later migrations, and this area is simply the place that was influenced the least, which does make some sense geographically/historically...

So basically Maykop_Novosvobodnaya is ENF + CHG + the northern endpoint of an Iranian-related cline, and Georgians are largely descendent from them.

Alternatively Kartvelian and high G2a could both be coincidental, with G2a there being due to much later intrusions and/or random founder effects, but at least it deserves a closer look.

Blasonario Cremonese said...

@ Gaska

All female samples present some calls on Y chromosome even if they haven't got one. Try to process one randomly from a paper where that one sample is listed as female, and you'll find some calls also for Y-SNPs.

Gaska said...

@Blasonario cremonese- All female samples present some calls on Y chromosome even if they haven't got one. Try to process one randomly from a paper where that one sample is listed as female, and you'll find some calls also for Y-SNPs.

Maybe you should explain it to me better, because my genetic knowledge tells me that women are XX and men are XY. Where is the Y chromosome in women?. Are you saying that there are YSnp calls on the X chromosome? Women can have the same Y-SNp calls as men? Is the first time I hear it. There are cases of XXY (Klinefelter syndrome that affects somemen).However, I have read that in some cases men seem to transmit mitochondrial haplogroups to their descendants.

What they have to do is check the snp of that sample, because the only thing they do is create confusion.

Blasonario Cremonese said...

@ Gaska

I can't explain anything, because I'm not a scientist nor a geneticist. I talk thanks to experience: if you process a BAM file of an ancient female sample through a software (i.e. BAM analysis kit), you'll find also some calls on Y-SNPs. I don't know why, but that's the actual situation.

So, if you give faith to all the Y-SNPs calls you find on samples, so you'll jump to the conclusion that all the samples we have until today are from males only.

Common sense taught to me that if you have a very big BAM file (so, complete and with many informations) and you get very few Y-SNPs results, so the sample of that BAM file was a female.

Another matter is that of multiple calls on different levels of different haplogroups, but that's not the argument.

Gaska said...

@B Cremonese-

Ok if you process a BAM file, they all have Y-SNP calls, regardless of whether they are women or men. I can not understand how a woman can have YSNP calls if she does not have Y chromosome. Someone will explain it to me. The Klinefelter syndrome (XXY) could be an explanation of women (XX) with a Y chromosome, but that is not correct because they are men with an anomalous duplication of their X chromosome.

So what is the difference between the case of Ireland that Arza mentioned and the rest of cases of HapY-R1b in ancient samples?
Why that case is a woman and the rest of the samples are not?
At least 4 SNP calls of R1b in a woman is normal ?. Have you seen any similar case?

Someone will have to explain it carefully. Could it be that the dating is very old and that complicates the explanation of the origin of R1b? I may be too suspicious but I'm sure someone has a convincing explanation.

Matt said...

Dates should be around as follows (from supplement table 1 of "Radiocarbon dates and Bayesian modeling support maritime diffusion model for megaliths in Europe" - https://www.pnas.org/content/116/9/3460/tab-figures-data and taking radiocarbon age as BP):

Primrose Grange, County Sligo, Ireland - 5230BP-4360 BP, median 4892 BP (3230 BCE - 2360 BCE, median 2892 BCE)

Ansarve Hage, Gotland, Sweden - 4595-1515 BP, median 3962 BP (2595 BCE - 500 AD, median 1962 BCE)

Carrowmore, County Sligo, Ireland - 6500BP-1500BP, median 4742 BP (4500 BCE - 500 AD, median 4742 BCE)

Kolin (???) - probably - https://www.megalithic.co.uk/article.php?sid=24394

Some date intervals are wide so reading paper will def. be needed to clarify. Samples in ENA are mostly labelled with Primrose (6x Ansarve, 1x Bal(?), 1x Carrowmore, 2x Kolin, 1x Lai(?), 2xMid (?), 12x Primrose, though that's 25, not 24).

When the paper is published it will be here - https://www.pnas.org/content/early/2019/04/09/1818037116

Blasonario Cremonese said...

@ Gaska

I've already explained to you the difference between that Irish sample and other surely R1b samples: that sample quoted by Arza has few calls on Y-SNPs and a huge BAM file (many GB), so it was a female.

Many samples, males and females have calls on R-M269 level, but they are mainly bad calls, because they also present nothing before and after M269 level. It's like to play domino: you can put down every card until that you want only if they are one-tied-to-another. So, if you want to be sure that M269 is correct, so you have to find positive calls also upstream.

I don't understand why you are so astonished by this: try yourself to process a female BAM file, and see what you gain when you set the BAM analysis kit only on Y chromosome.

Gaska said...


There are some cases in which the assignment of a haplogroup is very doubtful because they only find a SNp related to the haplogroup in question. In fact I have also seen some controversial cases in Spain. But there are some Snps that only occur in certain haplogroups and that are unambiguous when determining them.

How I will not be astonished, if women have R1b-M269 calls and they do not have Y chromosome. You said yourself that you could not explain it.

Blasonario Cremonese said...

@ Gaska

Please, don't quote me uncorrectly or incompletely: I said that I can't explain that because I'm not a scientist/biologist/chemist/geneticist/physician, in a word, I'm not a professional worker in the field.

Surely some SNPs occur only in certain haplogroups, but if you find only one SNP call positive for, take it for an example, M269 (that occurs only in M269), and nothing before and after it, not a single one or, maybe, one on L15 and another on M459, then how can you assign surely that sample? You can't.

Sorry, but it seems like you were born yesterday: if you try to process some files, like many users do sacrifying their free time to do something for others for free, then you would know something more about difficulties in this operation of processing BAM files.

I bring you an example: Genetiker, who has a little bit of bias, but concerning Y-SNPs calls he is an actual gentleman and very correct, had R1b calls on a sample from Monte Canelas. However, he had also H2 calls. Why, at the end, he concluded it was H2? Because that sample had the full complete chain of positive calls from A to H2, without any hole. On the contrary, to reach R-M269 that sample had a lot of holes of negative calls. Pay attention: not no calls, but negative calls.

Arza said...

@ Gaska
Female samples can have Y-DNA calls as random reads can be erroneously mapped to e.g. Y-chromosome giving false positives.

@ Blasonario Cremonese
But rather not in this case. Plenty of reads mapped to chrY (BWA default settings, hg19).

Primrose9BE - full BAM 3.0 GB, chrY only 31.2 MB, 1%
Primrose9WGC - full BAM 222.5 MB, chrY only 7.4 MB, 3.33%

Now I'm waiting for variant calls.

It can be contamination, wrong datation, some early Beaker buried in a reused tomb, but it's rather not a woman.

Gaska said...

@ B. Cremonese said- "I can't explain anything because I'm not a scientist nor a geneticist"

Ok I just remember you, what you said.

"if you try to process some files, like many users do sacrifying their free time to do something for others for free, then you would know something more about difficulties in this operation of processing BAM files"

Do not have any doubt that if I had time I would do it, because I understand that it has to be a very interesting job. It seems very good that people do it, it's true that this generates confusion, but also an interesting scientific debate. But obviously, the most credible opinions will always be those of professional geneticists.

"Genetiker, who has a little bit of bias, but concerning Y-SNPs calls he is an actual gentleman and very correct, had R1b calls on a sample from Monte Canelas"

I know the case, and I know Gnetiker also said that Atp3 was R1b-M269 because he found SNP-PF6518 in the BAM-file. We try to solve the controversy in Spain, and the truth is that although most of the geneticists we have consulted believe that it is r1b-M269, others think that it is not so. So who do we have to believe? We have no choice but to keep waiting.

There are some people (geneticists, archaeologists, amateurs, linguists) who have become the guardians of a certain genetic orthodoxy, especially in relation to R1b. It seems to me very well that everyone defends what they think, but always respecting what others think, and of course in the case of R1b which is fundamental to understand the genetics of Western Europe, dominant thinking can never impose their theories if they do not have conclusive and definitive evidence to do so. Meanwhile, we will be discussing all the time.

Matt has kindly given us the dating of the samples of Ireland, we will see what happens when the paper is published, because I guess the geneticists who have collaborated will have their opinion on whether the case in question is male or female, and about the calls downstream R1b-M269.


Gaska said...

@Arza- You said random reads can be erroneously mapped to e.g. Y-chromosome giving false positives.

The keywords are erroneously and false. That is to say, understandable difficulties in the interpretation of BAm files

"But rather not in this case. Plenty of reads mapped to chrY (BWA default settings, hg19)"

And Matt said us that Primrose County Sligo-2.892 BC. It could definitely become an interesting discussion.

Blasonario Cremonese said...

@ Arza

Plenty of reads mapped on Y chromosome? So not only at R1b level, I suppose. I haven't processed the file yet, because I saw that Richard Rocca did Primrose9BE just few minutes ago.

However, I see that the file, though big, has only a tiny percentage of material for Y chromosome.

Arza said...

Or maybe R.Rocca just made a mistake, as now he posted that:
prs009BE is as follows:

I2a2a1-CTS616
I2a2a1a-FGC15073
I2a2a1a1a-L1195


...and prs009BE is Primrose9BE.

Matt said...

@Arza, that would make sense. My guess if the calls were correct, it would probably be a late sample and possibly there'd be a good likelihood of some continuation of use of the megaliths by incoming R1b-M269 group.

capra internetensis said...

I guess no one is processing autosomal data? If he's EEF that will be interesting, if BB not so much.

OsoDanes said...

Linguistically speaking there is a decent set of comparanda between the Indo-European and Kartvelian proto-languages. I treat some of the better in my MA thesis (2017), some of which have gotten a neat picture and can be seen on my blog: http://loanwords.prehistoricmap.com/?s=Kartvelian. Since it only partially covers the data from the thesis, I should mention that also two verbs meaning to 'craft' or 'weave' show promising similarities, escpecially as these are items likely to be borrowed (pp. 66 and 111 in the thesis with proper references if of interest).

As linguists were obviously very limited by the size of the surviving languages, Kartvelian presenting an extremely confined geographical area and very few and contiguous languages (4!), so it is, of course, a very interesting discussion you got going on!

Romulus said...

@OsoDanes

How does Kartvelian compare to Uralic and Basque?

Gaska said...

@ Matt- "My guess if the calls were correct, it would probably be a late sample and possibly there'd be a good likelihood of some continuation of use of the megaliths by incoming R1b-M269 group"

@ Capra Internetensis- If he's EEF that will be interesting, if BB not so much.

IF the data given by Matt of 2,892 BC is true, it is totally impossible that it is a BB in Ireland, and that it has reused a megalith, it can only be a Neolithic farmer. And IF this is, so obviously and definitely, Delenda est Yamnaya.

Now it is going to be that the Basques are descendants of the Irish, of course although I have nothing against the steppes, I prefer the Irish.

OsoDanes said...

@ Romulus

They are doubtful. I have gleaned a very old attempt at connecting Uralic with Kartvelian through Altaic, but this does not appear to have been seriously entertained for more than a hundred years. The Basque connection lives in infamy in the heart and soul of many Caucasian nations (I had a hard time making an Armenian abandon the idea), but the evidence is very flimsy, see https://www.languagesoftheworld.info/russia-ukraine-and-the-caucasus/georgian-language-related-basque-another-european-outlier.html
Through happenstance I happen to own Bouda's 'Baskisch-kaukasische Etymologien', which, ironically, is a very good argument AGAINST the idea, picking at random from the numerous extant languages of the Caucasus (as Pereltsvaig will tell in the blog I referenced).

Gaska said...


I think the case of Neolithic R1b-M269 in Ireland is Prs002 BE, not PRS009 .. and it effectively has CTS3794, PFG501 L777 and CTS329. So everything will depend on the dating.

Some guardians of Kurganist orthodoxy in anthrogenica, have to be sweating and praying for the dating to be contemporaneous with the BBC. Whatever happens, it is clear that sooner or later R1b M269/L51 will appear in the West.

capra internetensis said...

@Gaska

Depends whether that date is on this particular skeleton. I'm just going to wait for the paper, I assume it will be out shortly, and I'm like five papers behind as it is.

Arza said...

R.Rocca@AG said...

Yes, sorry for the confusion. It was prs002BE, not prs009BE that had M269 level calls.

Oh.

*cancels all jobs queued on a server*

Gaska said...

Goran Burenhult-"In the nearby Primrose Grange area, about two kilometers southwest of Carrowmore, a large (c. 12 x 4 meters) megalithic stone-cist is currently being excavated by the Swedish team. The tomb has no remains of a court or a cairn (fig. 6) (Ó Nualláin 1979, Archaeological Survey of Ireland, SL014-166:62). Morphologically, the Primrose Grange tomb also lacks features that characterize the Carrowmore tombs (e.g. boulder circles, dolmens or small cists). Yet the ongoing excavation has shown that the tomb was in use at the same time as the Carrowmore cemetery. A radiocarbon (AMS) sample from the intact deposition layer inside the chamber has produced a date of c. 3915 cal. BC (3190±65 bc), and, thus, the date of tomb construction can be expected to pre-date that sample. Two more samples from burials in the chamber have given c. 3000 and 3500 cal. BC respectively (2410±80 bc and 2695±70 bc). The burials found in Primrose Grange Tomb 1 are almost all inhumations, and very few cremated human bones have been found.

Dragos said...

@ Aram

“sample": "Armenian:Average",
"fit": 2.3507,
"Kura-Araxes_Kaps": 52.5,
"Anatolia_MLBA": 47.5,
"Balkans_ChL": 0”

That’s not a very ideal fit. For ex I got twice as good with much older samples ..

Hugh Capet said...

@Davidski

Thank you for the reply. I agree with your assessment here.

JuanRivera said...

The radiocarbon dates of the later samples seem to fall within the period of Corded Ware (~2900 BC-~2350 BC) and Early Bell Beakers (from ~2800 BC onwards).

AWood said...

@Gaska

And a single sample proves what? That UK and Ireland weren't replaced with R1b during the BB period? Last I checked, every other sample was some form of I2 (if not all samples). You don't see a trend here? Seems you're the only one holding onto something.

Matt said...

@Arza, btw, I noticed your comment upthread: G25 coordinates, but from the 10-dimensional PCA published in the original paper it looks like PinarBasi is virtually identical to LBK_N, so there is a possibility that EEF were already WHG-shifted when they were entering Europe.

But then how does that fit with Koros_N being less WHG shifted than LBK_N? (Or even Barcin?) If you go by Lipson's study (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5973800/), just as one example, LBK about 5%, OK (looks fairly similar to G25?), but Koros have nothing beyond Barcin. Koros also earliest Neolithic barring the Starcevo set (who have only 2.5% HG).

JuanRivera said...

As said before, the sample may fall within Corded Ware and Beaker periods, and so it may harbor steppe ancestry.

Gaska said...


@Awood

Alright Aaron I guess you'll keep looking for P312 in the steppes and the way they came to Central Europe. If the dates of Ireland are confirmed, it is the nail in the coffin for the theory of the Kurgans. First Maykop then Iberia and now Ireland. Greetings to my friends from Anthrogenica, I guess they will be very happy. You still have more surprises especially referred to Df27

I still remember the great theories of the Kurgan BBs and the Similarity of the European prehistoric pottery, I suppose you will be satisfied, the international scientific community owe you a recognition.

JuanRivera said...

First, the most diversity of R1 as a whole is in a region spanning from the Baltic to the steppe and NW asia. Second, an auDNA signal corresponding to cultures in the steppe can be seen. Third, most of the Iberian R1bs and the Maykop R1a(s) have that steppe signal. It's not parsimonious to suggest that R1a (which I think will never be seen in pre-BA W Europe) and R1b are W European

Gaska said...

@Juan Rivera said---"As said before, the sample may fall within Corded Ware and Beaker periods, and so it may harbor steppe ancestry"

I'm sorry, Juan, but you should recognize that you have no idea about chalcolithic in Europe. Why you do not read any specific paper from Primrose site and then you tell us the relationship that the CWC has with it? It will be interesting to know your opinion. And remember that so far CWC is R1a and I2a, there is NO R1b, I do not think it is so difficult to understand. Regarding BB culture in Ireland, I recommend you read Firzpatrick and Cassidy.

Sooner or later cases of R1b-M269 will continue to appear in France and Spain. The later you acknowledge it, the greater the fall will be.

JuanRivera said...

There's a polish CWC R1b sample (plus another polish CWC sample which is doubtfully R1b). Plus, about the W European Chalcolithic, being "local Calcholithic" doesn't prevent from having steppe ancestry, as with Iberia_CA_stp.

JuanRivera said...

RISE1, the confirmed CWC R1b sample, is dated to ~2700 BC and has K1b1a1 as mtDNA.

EurDNA said...

Starcevo - Koros- Cris are essentially a unit, and would form a clade with Neolithic Anatolians
I would be very surprised if Epipaleolithic Pinarbasi would be a unit with those, despite what the paper suggests

JuanRivera said...

RISE1's indeed of low quality. However, it seems to be truly R1b. Here's the most detailed info: http://arshba.ru/ancient-western-eurasian-dna-of-the-copper-and-bronze-ages-t1147.html. The one that was changed to R1a was a german CWC sample.

JuanRivera said...

And I realize I made an error by speaking of another possible R1b CWC polish sample, whereas actually RISE1 is the only R1b CWC sample. There's also U106 in Swedish Battle Axe, which is a subset of CWC.

EurDNA said...

@ Juan
That Sample from Sweden is c. 2200 BC, which is Flint Dagger culture, not Battle Axe

JuanRivera said...

Thanks for correcting that.

Ric Hern said...

Regarding R1b. If it did not spread as a subset of Corded Ware there are still some R1bs in Germany Baalberge etc. that could point to some earlier R1b migration. The Tabiano Coloured Horse of Salzmünde +3100 BCE. with probable link to the Steppe or maybe even a closeby Steppe related population...There is still Usutovo and Cernavoda Cultural expansion towards the West to look at if Corded Ware doesn't fit the bill for R1b spread. So if nothing pops up from Western Ukraine, Eastern Poland or the Transylvanian Plateau then only will I start sweating...Heheheeh. The presence of U5b2a2 in Poland, Germany and eventually in Ireland, but nothing South of the Carpathians and Alps point to me at a Northern origin of R1b-L51. The earlier presence of MtDNA U5a1b in Ukraine, Latvia and eventually in Ireland and nothing South of the Czech Republic also points me to a Northern migration. So somewhere between the Elbe and the Dnieper North of the Carpathians and Alps we will find the answer.

Ric Hern said...

If RISE1 was indeed R1b it shows some contact between R1b and GAC. K1b1a1 was found in LBK in Austria and GAC. So clearly a Farmer related expansion of this MtDNA...Contact between GAC and R1b points to yet another possibility of R1b-L51 in the North...So we now see two Corded Ware related MtDNA Haplogroups possibly connected to R1b. U5a1b and K1b1a1....

Ric Hern said...

So currently I see a migration of R1b between the Warta and Notec Rivers towards the West...So either down the Warta or down the Vistula and Notec from Western Ukraine and or Southeast Poland.

Gaska said...

@Ric

If all you have are two mitochondrial haplogroups to continue maintaining an eastern origin of L51 and a migration of R1b linked to the CWC, you will end up defending that our origin is on Mars. Anything to keep denying the evidence that P312 is absolutely western and that the theory of the Kurgans is history. You should check the mitochondrial databases because K1b1/a1

I0407-Iberia-Dolmen de la Mina-3.750 BC
I1272-Iberia, El Mirador, Atapuerca, Burgos-2.676 BC
Iberia-El Juncal- BB culture 2.651 BC
I6466-Iberia-Cueva Verdelha, Vialonga, Portugal-BB culture 2.500 BC
MA89-Sardinia, Philigosa-3.190 BC
I2637/I2978-England, Neolíthic-3.499 BC/3.179 BC
I2416-Amesbury-Boscombe Bowmen, BB culture-2.303 B
I2446-Yarnton-2.297 BC
I2566-Amesbury-2.119 BC
I4951Flying School-2.150 BC
I2441-Poland, Kierzkowo, GAC-3.100 BC
I5530/I5019- Germany, Weichering, BB culture-2.250 BC
RISE98- Sweden, Neolithic-2.153 BC

GAC culture (3.400-2.800 BC)-That is, we have K1b1/1a in Iberia 350 years after the GAC existed, and you said that it is a CWC mit-Hap linked to the expansion of R1b ?????????

It is also one of the haplogroups I said that mixed with the R1b BBs that arrived at the islands. As you will see it was quite frequent in the British Neolithic, and the connection between Sardinia, Iberia and the British isles is evident.

Ric Hern said...

@ Gaska

It seems that you forgot (maybe conveniently) the LBK sample from Austria which predates all those that you give as examples....And you seem not to see the R1bs +-5000 BCE in Ukraine accompanied by MtDNA Haplogroup U5a1b also found in Rathlin Ireland later...And if you read what I said I was also referring to K1b1a1 in GAC and Corded Ware....

Ric Hern said...

U5a1b was also found in Latvia in the Context of Corded Ware just like that K1b1a1 sample in Poland...

Gaska said...

There are approximately 50 mitochondrial haplogroups shared by the different BB regions in Europe. In other words, the BBC is an absolutely different people that migrated across Europe. In Central Europe they mixed with women from the CWC who, in turn, had some haplogroups from previous migrations from the steppes. That is the relationship that Western Europe has with the steppes. Nothing else, the rest is imagination of some geneticists and archaeologists who have found in Gimbutas an explanation that today does not hold, no matter how hard some insist.

The language? - We have already seen what has happened in Iberia.
Mitochondrial haplogroups?- 90% Western-Neolithic famers and HG
R1b? You do not even have L51 or P312 in the steppes.
BB culture?-Obvious origin in Iberia



Ric Hern said...

And U5b2a2 in Poland Corded Ware, Blatterhole and Rathlin Ireland. That all seems rather Northern to me...

Ric Hern said...

@ Gaska

Like I said, some samples from Western Ukraine and Eastern Poland will clear up this issue...

Samuel Andrews said...

Patrilocalism in Neolithic Europe....(as in Kurgan culture)

-A new abstract says in Megalithic burials in northwestern Europe show men from same paternal lineage across multiple generations in these burials.

-Someone apart of that upcoming study published a thesis on Megalithic burials in Gotaland. All the males belonged to the same Y DNA haplogroup: I2a1b1a.
-The only other Sweden Neolithic sample is I2a1b1a.


-All, Neolithic British males belong to I2a1b1 & I2a2a1a1a (one is I2a2a1b).

-All Globular Amphora males belong to I2a2 (all tested are confirmed I2a2a1b2).

-A large majority of Late Neolithic/Chalcolithic males from Basque country/northern Spain belong to I2a2a2a.

-Majority of males from Late Neolithic/Chalcolithic Portugal belong to I2a1a1-M26.

All over late Neolithic/Chalcolithic Europe, we're seeing patrilineal kin groups just like the ones in Kurgan cultures.

Ric Hern said...

@ Gaska

And you do not even have R1b that is closer to Z2103 (Brother of L51) in Iberia before 3000 BCE. So do not be so hasty...

Ric Hern said...

@ Samuel

Thanks. Very interesting. So it looks like I2a2a was the common Ancestor of Globular Amphora, Late Neolithic/Chalcolithic Spain and some Neolithic British...So maybe a split near Austria LBK....?

Ric Hern said...

And maybe a I2a1 split during the Earlier Neolithic along the Western Mediterranean ?

Davidski said...

@OsoDanes

Linguistically speaking there is a decent set of comparanda between the Indo-European and Kartvelian proto-languages. I treat some of the better in my MA thesis (2017), some of which have gotten a neat picture and can be seen on my blog: http://loanwords.prehistoricmap.com/?s=Kartvelian.

Thanks, interesting blog.

Samuel Andrews said...

@Ric Hern,

I2a is a Paleolithic lineage in Europe. Look at these age estimates....
https://www.yfull.com/tree/I2/

There's no common recent paternal ancestry between all these Neolithic I2a haplogroups. I2a2a1b, I2a2a2a, I2a1b1, I2a1a1 all existed in the Paleolithic. They were absorbed by new farmers from Anatolia, became popular in Late Neolithic.

Aram said...

Dragos

I see. Can You do now the same for modern Greek samples from Anatolia that are present in Globe25?

That Anatolia_MLBA can be a more complex place than just IE Anatolians and non IEs. The kingdom Ahhiyawa is attested in that period in western Anatolia. Most scholars now believe it was a Greek kingdom.
https://en.wikipedia.org/wiki/Achaeans_(Homer)#Hittite_documents


Bob Floy said...

The Kartvelian word for "Blood" is shared with both PIE and Sumerian?
Fascinating.

zardos said...

"There's no common recent paternal ancestry between all these Neolithic I2a haplogroups. I2a2a1b, I2a2a2a, I2a1b1, I2a1a1 all existed in the Paleolithic. They were absorbed by new farmers from Anatolia, became popular in Late Neolithic."

Became popular is an interesting description for the fact that the original, main early Neolithic lineages of G were largely annihilated in the process of related I lineages taking over.
I see some sort of common origin in HG clans taking over Neolithic communities, primarily the females, and expanding aggressively in all directions.
They were more warlike than their Neolithic predecessors.

I doubt that process happened multiple times in all corners of the continent, independently but synchronous.

So where did the overtake start, which are the earlist dates for I2 lineages taking over?

Ric Hern said...

@ Davidski

Thanks, that is interesting. Mostly looks like Proto-Level contact between North Caucasus and the Steppe especially if some Proto-Indo-Europeans started to expand Westwards before Maykop really became notable...eg. the Horse Headed Scepter expansion predates Maykop if I understand correctly ?

zardos said...

The best candidate should be Northern Europe, were the HG were strongest and the otherwise well organised Neolithics weakest.
There Neolithics were pushed back, with Neolithic females being taken which produced pottery in a HG subsistence.
Then the acculturated HG lineages started to become warlike pastoralists which expanded southwards.
That corresponds well with the decrease of WHG ancestry from North to South.
But the male lineages were kept and foreign males largely eliminated.
They did the same steppe people from the East did, probably they even taught them via GAC.
But the steppe people gained an edge over them and did the same they did before, just more successfully and lasting.

The Megalithic culture along the Atlantic coast is worth to look at.

For the scenario to work the older dates for I2 takeovers should be earlier in the North.

Davidski said...

@zardos

What if local forager males were incorporated into farming societies throughout Europe because they had specific hunting and foraging knowledge and skills that came in useful when climate change and/or the over exploitation of farming lands made farming less productive at some point?

It's possible that the farming societies that took advantage of this process could then expand into the territories of those that didn't, hence the expansion of genetically related populations rich in hunter-gatherer ancestry, such as those associated with the TRB and Globular Amphora cultures.

Samuel Andrews said...

Look at difference between Poland_TRB:N19 and Poland_TRB:N18. N19 is identical to contemporary Hungary farmers. N18 is most akin to British Neolithic, France Neolithic.

It's direct proof of migration in Neolithic & the diversity of the Funnel Beaker culture. N18 is clearly descended from western European migration, N19 from local Danube/Lengyal farmers.

The high level of HG ancestry in Funnel Beaker & Globular is in part due to western European farmer ancestry. But, they also clearly have dose of ANE-infected eastern European hunter gatherer stuff.

Which is where Globular Amphora inherited I2a2a1b2 (which was popular in eastern European hunter gatherers). But, I2a1b1 in Swedish Funnel beaker is probably from western Europe. It was common in Britain & pops up several times in Iberia & France.

EurDNA said...

As has been pointed out by Sam, the bounce-back phenomenon entailed deeply divergent I2 lineages, therefore there's unlikely to be a single, overiding explanation.
Not to focus on blood & gore, Michelsberg would be interesting. The preliminary mtDNA study suggested differentiation between the sacrificed and the not. However, we have to see if the distinction is maintained when the full-genome version is published

Gaska said...

@Ric Hern- "And you do not even have R1b that is closer to Z2103 (Brother of L51) in Iberia before 3000 BCE. So do not be so hasty..."

Come on Ric, why we don't all try to forget our preferences and try to find a reasonable explanation for the issue. It is absurd to always discuss the same thing, there are still some reasonable explanations, and the issue of Ireland is going to depend on the dates that the paper tells us. I think the ones Paulsson mentions are the same ones Swedish archaeologists did years ago. Maybe they have new dating of the bones and obviously it could be a reuse of the megalith. However, all the indications are that a thorough analysis of French, Italian, Iberian and even Irish deposits will be definitive to clarify the issue.

I guess a lot of people will have panicked, but they've already been on the edge of the cliff many times.

Samuel Andrews said...

Or, if you look at ILK002 & ILK003 from Globular in Ukraine, they're of mostly western European farmer origin, showing a far-reaching population movement in Neolithic European. They weren't anymore native to Ukraine than Yamnaya was.

zardos said...

What you propose did happen during the initial colonisation resulting in single I2 individuals here and there, but no big change in the colonisers culture.

What happened in the North is fundamentally different, because the HG lineages were able to fight the Neolithics back, stole their women and began to adapt to the new tools on their own.
That resulted in a warlike and pastoralist new culture which spread much beyond the region were it was needed for survival because they were more competitive.

That spread with increasing WHG ancestry, l2 male lineages and new economic and military strategies didnt happen because of peaceful cooperation.

They gave more local males a chance than later steppe Beakers it seems.
Still we can speak of a conquest and mass killings for sure.

It just went more under the radar so far because the genetic and cultural impact was less significant compared to the steppe expansion.

But little of the WHG increase after the first Neolithic wave was local I would guess, but the result of a new conquest from HG lineage derived groups from Northern Europe.

I wouldnt exclude the same thing happen more than once, after all steppe people are just the same scenario further East.
But they too were independent HG adopting by themselves.

HG prefer to be raiding warriors and pastoralists, before doing field work.
You see something similar in North American Indians and the Mongol transition too.

Richard Rocca said...

All, sorry for the confusion. Serves me right for trying to post late while rushing to go to dinner. It was prs002BE, not prs009BE that had M269 level calls. Not only does this sample have M269 level calls, but most calls leading to the M269 branch are indeed positive. I am using the 2017 ISOGG tree for my calls:

A-V221 >> Positive for equivalent Z11905
BT-M42 >> Positive for equivalents Z40399, M9115, P97, M9151, M9177, M9301, M8973
CT-M168 >> Positive for equivalents M5655, M5695, CTS7257, M5780, M5811, Z40572, M5590
CF-P143
F-M89 >> Positive for P141 equivalent
IJK-L15
IJ-P124
K-M9
K(xLT)-M526
K-YSC0000186
P-P295 >> Positive for equivalents M1267, CTS12524, PF5481. Negative for equivalents F1660, M1258
R-M207 >> Positive for M207
R-M173
R-M343
R-L754 >> Positive for CTS3794
R-L389
R-P297 >> Positive for PF6501
R-M269 >> Positive for L777, CTS329


Positive haplogroup I SNPs CTS2514, CTS8963
Positive for haplogroup N SNP CTS7885
Positive for haplogroup C2b1a2b1b SNP M4620

I have put the calls here for folks that want to investigate further:

https://docs.google.com/spreadsheets/d/1LemAdMuBsPWWM9FwZh2sPLmHk25KsW0NRJhoZOpOakw/edit?usp=sharing

zardos said...

EurDNA
"As has been pointed out by Sam, the bounce-back phenomenon entailed deeply divergent I2 lineages, therefore there's unlikely to be a single, overiding explanation."

If they stem from the same or closely related HG lineages, they can be different parts of the same phenomenon.
Its just that they expanded in different directions so to say.

To say the replacement of the original Neolithic colonisators did happen throughout Europe randomly,with the same end result everywhere, is putting random behaviour and chance on a new and rather unlikely level.

Look at the yDNA in Iberia,actually you have two near complete male lineage replacements.
First I2 taking out a G-E-I mixed population, then R1b annihilating I2.

Same in Britain, France, Central Europe...

R1a and R1b are deeply diverged,hopped on the same train and still you end up with provinces of one particular R1b variant.

The same just happened before already with I2-WHG expanding in all directions.

Gaska said...


And why nobody can think that L21 originated in Ireland instead of in mainland europe?

Andrzejewski said...

@Samuel @Zardos @Davidski @Ric "-All Globular Amphora males belong to I2a2 (all tested are confirmed I2a2a1b2)."

OK. We all agree that for some reasons there was a bounceback of WHG/SHG ancestry during Late Neolithic period, and that it was very pervasive, even to the point of extermination of the overwhelming majority of Anatolian G markers. My theory is that because of that plague which farmers were more vulnerable to (similar to Spaniards v. Aztecs), WHG gained the advantage and pillaged farmer women, killing surviving males following the already devastating plague (and ensuing famine).

What I DON'T understand is:

1. How come GAC, which was the main substrate of the CWC (25% which wasn't from Kurgan/Ukraine Eneolithic) was rich in WHG, let alone Bell Beakers in Northwestern Europe mingled with local females to reduce Steppe aDNA component from 75% to a further 40%-50%, and yet - supposedly that Western/NW population was much richer in WHG ancestry but modern Europeans have no more than 15% WHG in them. What can account for that? (Sam even pointed out that the WHG-rich samples in TRB/GAC came from the west, so logic dictates that WHG ratio should increase rather than DECREASE)?

2. Moreover, even in Scandinavia where I2a is strong and which allegedly was the focal point were HG bands started reclaiming farmer territory and expand Southbound - WHG are no more than 20% at best (excluding Finland!) while EEF are at least 30%?

3. Does WHG go underestimated and flies under the radar because Steppe groups were also rich in U5b and I2a, so researchers overestimate Steppe contribution in lieu of WHG/SHG?

4. Can it be that Globular Amphora and later on Corded Ware were blond, blue eyed and fair skinned because of the SHG admixture rich in ANE (AG3 like)?

zardos said...

The WHG expansion as I see it came from heavily diluted male lineages which took local females as they advanced and were at no point in time, since they transitioned to the productive economy, more than 50 percent WHG.
Even on the contrary, autosomally the Neolithic part was always dominant and increased just due to their expansion.

Andrzejewski said...

@Samuel "It's direct proof of migration in Neolithic & the diversity of the Funnel Beaker culture. N18 is clearly descended from western European migration, N19 from local Danube/Lengyal farmers.

The high level of HG ancestry in Funnel Beaker & Globular is in part due to western European farmer ancestry. But, they also clearly have dose of ANE-infected eastern European hunter gatherer stuff.

Which is where Globular Amphora inherited I2a2a1b2 (which was popular in eastern European hunter gatherers). But, I2a1b1 in Swedish Funnel beaker is probably from western Europe. It was common in Britain & pops up several times in Iberia & France."

Which should only corroborate that WHG ratio should increase instead of decrease from CWC as we move North and West into territories where foragers are strongest and famers weakest. But it apparently didn't turn out that way, did it?

JuanRivera said...

It is me or do CSID (Congenital Sucrase-Isomaltase Deficiency) rates show a correlation with levels of ANE ancestry?

Andrzejewski said...

https://en.wikipedia.org/wiki/Den%C3%A9%E2%80%93Yeniseian_languages

If linguists like Vaida find strong connection between Ket (Yenisseyan) and Native American languages from 24,000ybp and they are both Ancient North Eurasian languages from the time of MA1, then I don't see any reason NOT to relate some early putative form of PIE with either Yenisseyan or another EHG/ANE language on the one hand, or to Kartvelian/Northwest Caucasian languages on its CHG side (although I strongly suspect that both Kartvelian and NW Caucasus languages arrived with Anatolian farmers, and NOT with autochthonous CHG).

John Thomas said...

OT

David,

I'm sure you're aware of that Nature published study, released today, of neolithic British DNA and haplotypes.

Matt said...

@John Thomas indeed (basically a copy of Brace's study on biorxiv, but looks a bit more cleaned up and professional).

Haplotype donation from moderns to form ancients: https://imgur.com/a/rw1m1jG

However, no GAC or Hungarian Copper Age, which would be much more interesting.

JuanRivera said...

As for Ireland, it has been somewhat of a genetic sink until the last ~1000 years. As such, it's not parsimonious to assume Ireland as the place of origin of L21.

Matt said...

One of the improvements in Brace's paper seems to be a bit more references and discussion around their pigmentation genetics work, and a link to the Hirisplex reference paper Chaitanya 2018 (https://www.sciencedirect.com/science/article/abs/pii/S1872497318302205), which they use.

For anyone who wants to put these into context (excerpts from paper with reference on what those *mean* from the Hirisplex reference paper): https://imgur.com/a/w4frmoJ

I still think the issues that Iain Mathieson tweet about that prediction is pretty questionable on ancient individuals still stand tho.

Samuel Andrews said...

Thanks for sharing new Hirisplex Matt. Agreed, it's hard to know how accurate ancient DNA results are.

Samuel Andrews said...

But, it would be interesting to see new results for ancient pops using this new Hirisplex.

MaxT said...

@1Andrzejewski

That reminds me of Mark Pagal et al 2013 study. He estimates 15,000 year ago, many languages spoken today in Eurasian had common ancestor.

"If you went back 15,000 years and spoke these words to hunter-gatherers in Asia in any one of hundreds of modern languages, there is a chance they would understand at least some of what you were saying.That’s because all of the nouns, verbs, adjectives and adverbs in the four sentences are words that have descended largely unchanged from a language that died out as the glaciers retreated at the end of the last Ice Age. Those few words mean the same thing, and sound almost the same, as they did then.The existence of the long-lived words suggests there was a “proto-Eurasiatic” language that was the common ancestor to about 700 contemporary languages that are the native tongues of more than half the world’s people."

https://cdn.zmescience.com/wp-content/uploads/2013/05/language_families.jpg

OsoDanes said...

@Rob Floy
The comparanda are not necessarily related. They’re grouped simply because they have been proposed as related to the same PIE root. Sumerian data is generally very difficult to work with from a comparativist POV as isolates give away very little in terms of diachronic perspective. There are links on the blog / in the thesis to pursue the arguments further.

Romulus said...

The new paper really didn't tell us anything we didn't already know about British Isles Neolithic DNA, the Mesolithic data is new, great to finally see Cheddar Man's y-hg (i2a2).

Dragos said...

What happened to “C for Cheddar “?

Link, anyone. ?

Matt said...

For those that wish, that paper on ancient dna of megalithic burials is now published - https://www.pnas.org/content/early/2019/04/09/1818037116

New article on the same - https://phys.org/news/2019-04-megalith-tombs-family-graves-european.html

capra internetensis said...

Primrose 2, 3790-3660 cal BC, female - was that the one people were looking at?

Richard Rocca said...

capra internetensis said...Primrose 2, 3790-3660 cal BC, female - was that the one people were looking at?

That's the one. I had written on my first post about it that it was likely a female due to its low Y-DNA to total DNA ratio. Good to see they confirmed it.

«Oldest ‹Older   1 – 200 of 286   Newer› Newest»