search this blog

Sunday, May 21, 2017

Steppe invaders in the Bronze Age Balkans


In a recent blog post announcing the end of the search for the late Proto-Indo-European (PIE) homeland I wrote this:

But of course I2a has also been recorded in prehistoric samples from the Pontic-Caspian steppe. So, you might ask, why did the populations migrating out of the steppe belong to R1a and R1b, and why did some of them seemingly carry only R1a while others only R1b? This can be explained by local founder effects on the steppe due to patrilocality. Moreover, it's possible that some groups moving out of the steppe did carry high frequencies of I2a, but they're yet to enter the ancient DNA record.

Actually, in hindsight, such a population has probably already shown up in the ancient DNA record, via two Early Bronze Age (EBA) individuals from the Balkans in the Mathieson et al. 2017 preprint:

Balkans_BronzeAge I2165: Y-hg I2a2a1b1b mt-hg T2f 3020-2895 calBCE

Yamnaya_Bulgaria Bul4: Y-hg I2a2a1b1b mt-hg ? 3012-2900 calBCE

Both samples are from burial sites in present-day South-Central Bulgaria. Apart from sharing I2a2a1b1b, they each pack a fair bit of Yamnaya-related ancestry and are dated to a very similar time period. Unlike Bul4, I2165 does not make the cut archaeologically as a Yamnaya sample, but he does come from a Tumulus (Kurgan-like) burial, so perhaps he's from a group influenced by Yamnaya?

By the way, the I2a2a1b1b lineage is also shared by Yamnaya_Kalmykia RISE552, and as far as I can tell, the oldest individual sampled to date belonging to this line is Ukraine_Neolithic I1738, dated to 5473-5326 calBCE. So I2a2a1b1b appears to be a Pontic-Caspian steppe marker.

The same paper also includes the following individual from present-day Bulgaria dated to the start of the Late Bronze Age (LBA), which is roughly when the Mycenaeans appeared nearby in what is now Greece:

Bulgaria_MLBA I2163: Y-hg R1a1a1b2 mt-hg U5a2 1750-1625 calBCE

This guy is the most Yamnaya-like of all of the Balkan samples in Mathieson et al. 2017, and, as far as I can see based on his overall genome-wide results, probably indistinguishable from the contemporaneous Srubnaya people of the Pontic-Caspian steppe. He also belongs to Y-haplogroup R1a-Z93, which is a marker typical of Srubnaya and other closely related steppe groups such as Andronovo, Potapovka and Sintashta. So there's very little doubt that he's either a migrant or a recent descendant of migrants to the Balkans from the Pontic-Caspian steppe.

The presence of multiple individuals like this in the still rather spotty Balkan Bronze Age ancient DNA record suggests that this part of Europe experienced sustained and possibly at times large scale incursions of various peoples from the Pontic-Caspian steppe throughout the Bronze Age.

Here's one of the Principal Component Analyses (PCA) plots from Mathieson et al. 2017, edited by me to highlight the above mentioned three samples, as well as the anything but weak impact of gene flow from the Pontic-Caspian steppe on the Balkans during the Bronze Age. Just in case some of you are confused, I added an arrow pointing to the cluster that most of the Balkan Bronze Age samples are pulling towards.


Of course, many of us are now eagerly awaiting a paper on the genetic origins of the Minoans and Mycenaeans. The latter are one of the few attested Indo-European speakers from prehistory, so their genetic structure may prove pivotal in the Indo-European homeland debate.

I know for a fact that a couple of ancient DNA labs have been working on such a paper for a while now, but haven't heard anything about the results. However, just looking at the PCA above, I'd be shocked if the Mycenaean samples did not show a strong signal of gene flow from the Pontic-Caspian steppe. If so, the implications of this will be obvious.

Reference...

Mathieson et al., The Genomic History Of Southeastern Europe, bioRxiv, Posted May 9, 2017, doi: https://doi.org/10.1101/135616

See also...

Main candidates for the precursors of the proto-Greeks in the ancient DNA record to date

Steppe admixture in Mycenaeans, lots of Caucasus admixture already in Minoans (Lazaridis et al. 2017)

281 comments:

«Oldest   ‹Older   201 – 281 of 281
Arza said...

@ EastPole

When I saw this line I immediately recalled "your" lines. Maybe indeed there was something interesting happening in line with this... line. Baltic_BA could be simply shifted towards Latvian_HG in this case.

Interesting citations BTW, especially the one about "laryngeals".
____
Demo of the spreadsheet with Malak_Preslavets regression line.
https://3.bp.blogspot.com/-LVT291QpQwQ/WSWr1EV2E_I/AAAAAAAAALc/Aatv4-BRwlYkb6aE8WnBG7GW5H3kIjtZgCLcB/s1600/malak_preslavets.png

Gioiello said...

@ EastPole
"Your ignorance is probably the result of poor quality pseudo-scientific books you have been reading".
Goth mimz
Lith mesa
OCS meso
Alb mish
Arm mis
Skt mas /masa-
Toch B misa
Lat membrum (dick)
Grk menigks
OIr mir
Rus mjazdrà

Davidski, have I to laugh? Everything from Polish?

EastPole said...

@Gioiello
“Davidski, have I to laugh? Everything from Polish?”

Western Slavic languages like Polish are more archaic and conservative than Southern Slavic like OCS or Eastern Slavic like Russian, not to mention Albanian, Latin or Germanic.

Comparing Slavic languages with Vedic Sanskrit, Avestan, Homeric Greek, Gothic and others which were influenced by CWC migrations into Asia, Balkans, Scandinavia etc. we can reconstruct many CWC words.

So as explained in the examples given above Polish ‘mięso’(mienso) was probably how CWC people pronounced “meat” and Polish 'velna' was how they pronounce 'wool'.

Notice that CWC originated and expanded from Poland, was R1a dominated, produced populations autosomaly close to Slavs like Sintashta or Andronovo, maybe this is the reason.

Alogo said...

Davidski said: "I think that the horse/chariot complex had a profound impact on Mycenaean culture, but that it was a secondary layer of steppe influence in the Mycenaeans."

That's very possible. Good, dense sampling from the early 3rd millennium down to the late 2nd might be able to conclusively solve some still open questions about the nature of interactions in that period.

Kuz'mina's monograph is very interesting but some of her arguments from mythology were kinda tenuous, to be honest.


@EastPole, stop confusing language with genetics and assuming cognates have Slavic roots, man.

Gioiello said...

@ EastPole

All what you said is wrong just from a linguistic point of view. Two phenomena (palatalization and polnoglasie: the last "The change has been dated to the second half of the eighth century, before any Slavic languages were recorded in writing") explain why what you said is false and with any knowledge of linguistics: ask who teaches at Goettingen.
Latin membrum demonstrates that a word *mems-ro- was in old Latin, and the vocal is -e-: no palatalization. The problem of the slides y/w is other thing, which has nothing to do with what you said.

Anthro Survey said...


@Davidski

(@Kurti--care to comment?)

Ok, I do see the Armenia Chalcolithic 20% EHG score in Laz's paper. The thing is, how do we know that most of that "blue" is actual EHG and not simply an affinity deal going on with non-EHG ANE & Villabruna?

CHG also has substantial "EHG" on that spread. Certainly cannot be EHG, but extra ANE and Villabruna-related dna in addition to the Eastern Basal-rich Iran component making up CHG.

Davidski said...

I haven't had time yet to make fun of this press release from Max Planck, but I might do so in the future.

http://www.shh.mpg.de/415581/genosicilygreece052017

We have already discussed the paper that the press release is about in another thread, so there's no point discussing it again here.

Davidski said...

@Anthro Survey

It appears to be something linked to early Yamnaya, but I haven't been able to characterize it properly.

https://drive.google.com/file/d/0B9o3EYTdM8lQLXZIdUpyV3RQWTA/view?usp=sharing

Anthro Survey said...

Hmmm, ok. Yes, more characterizing is in order. Methinks a combo of direct Yamna-related ancestry and non-Yamna ANE/Villabruna.

Awale from Anthromadness makes a good cautionary point about tree mixes here:
https://ibb.co/jqYKna

Rob said...

@Dave
3000 years boy
Evidence please ?

Rob said...

@ Dave

"It appears to be something linked to early Yamnaya, but I haven't been able to characterize it properly."

Yamnaya :3000 BC
Armenian Chalcolithic :4000 BC

Want to try again ?

Davidski said...

I2a2a1b1b is from the steppe and you know it.

Rob said...

Your weak with basic numbers and it' shows

Davidski said...

You have no argument because you haven't backed up any of your claims with evidence.

Davidski said...

I'll just keep re-posting my earlier comment until you choose to produce some evidence.

The Ukrainian is a better match, and is not only older, but also located between Bulgaria and Kalmykia. See this?

Yamnaya_Bulgaria Bul4 I2a2a1b1b
Ukraine_Neolithic I1738 I2a2a1b1b
Yamnaya_Kalmykia RISE552 I2a2a1b1b

Bulgaria_EBA I2175 I2a2a1b1

I keep showing the evidence, and you keep ignoring it.

Rob said...

It's clear you don't know anything about the material at hand, so I'll help you. The late coper Age I2a2 is in the rhodopes and from 3300BC, but had begun by at least 3500 BC. .
That's 500 years earlier than Yamnaya and several hundred mark kilometres away .
Did they use teleporters or are you going to invent the Neolithic steppe PiE migration theory ?

Rob said...

Sure come back when you have evidence and learn basic culture - history and chronology.

Davidski said...

Your reasoning makes no sense and you have no hope of winning this argument unless you post evidence backing your claims.

I'll just keep re-posting my earlier comment until you choose to produce some evidence.

The Ukrainian is a better match, and is not only older, but also located between Bulgaria and Kalmykia. See this?

Yamnaya_Bulgaria Bul4 I2a2a1b1b
Ukraine_Neolithic I1738 I2a2a1b1b
Yamnaya_Kalmykia RISE552 I2a2a1b1b

Bulgaria_EBA I2175 I2a2a1b1

I keep showing the evidence, and you keep ignoring it.

Rob said...

And please adress how Yamnaya could have impacted the formation of armenia Chalcolithic when it's over 1500 years later.
That's a new level of retard

Davidski said...

I said "it appears to be something linked to early Yamnaya".

Something linked to early Yamnaya can date to earlier than 4,000 BC.

Rob said...

Yeah it's called the Majkop migration (but shhh don't tell da Kurgan hypotenuse people because they will be upset)

Davidski said...

It looks like there was an expansion of an EHG-rich population both to Varna and Armenia at about the same time.

This is unlikely to be Maykop.

Rob said...

Yes there seems there was female exogamy from the steppe toward the developed civilisations on its flanks. They might have played an important role in cultural and linguistic transmission. Really quite expected.
It seems the model of the more nuanced scholars will be proven correct- the steppe was a good conduit / transmission zone, but not the homeland. Even the blind die-hards will have to accept the classic Kurgan hypothesis is incorrect.

Samuel Andrews said...

@Rob,

You remind me of a lot of people who post at places like forumbiodiversity, anthroscape, and anyother racist "anthro" or "Alt-Right" sites. Geneticker, German Dziebel, and political extremist(Nazis, Communists, Libertarians) are apart of the same disgusting family.

For the Nazi his enemy is the Jew, for you your enemy is the "Steppe crowd" aka people with brains.

Samuel Andrews said...

@Rob,
"Yes there seems there was female exogamy from the steppe toward the developed civilisations on its flanks. "

Yeah, that's why only 6% of Northern Europeans have U5a while 20% of Yamnaya had U5a. And that's also why so few LNBA Europeans had Y DNA G2a, I2a, or J but had 90%+ R1. Don't throw out bull shit about sex bias admixture if you don't know a damn thing about mtDNA.

And in case you forgot EEF ancestry didn't appear on the Steppe until Poltvaka_outlier. Groups like Yamnaya and Corded Ware formed without help from EEF civilizations.

Rob said...

@ Sam

You and brains don't go together in the same sentence.
Why don't you to back to your local elementary school, and learn how to spell and find Europe on a map.
As for your faux virtue signalling, if I am racist, how is it that I am propounding a West Asian homeland for PIE ? Doesn't quite make sense. No, the only people that are biased are lame *little* boys with big chips on their shoulders who need to live vicariously through some imagined whizz-bang steppe invasion. It is an interesting psychological analysis.
In fact, wasn't it you who has been blocked from fora for calling Oriental people "Mongies" ?

Davidski said...

@Rob

Yes there seems there was female exogamy from the steppe toward the developed civilisations on its flanks. They might have played an important role in cultural and linguistic transmission.

Unlikely, because...

Balkans_Chalcolithic_outlier I2181 Y-hg R mtDNA HV15

This individual is heavily Steppe admixed with Khvalynsk-like admixture, and belongs to Y-hg R but has an mtDNA that looks more Balkan farmer than Eneolithic Steppe.

Rob said...

@ Dave

You cherry pick analysis based on individuals you deem to fit your narrative. But all you're doing is exposing the fact that you probably didn't even take high school biology.
Otherwise, you'd be aware that - unless specific data (known admixutre & migration Hx) is known- the GW variation rapidly becomes de-coupled from uniparental markers.

Samuel Andrews said...

@Rob,
"if I am racist, how is it that I am propounding a West Asian homeland for PIE ? "

You're a special type of racist. This type is especially twisted. Clean wholesome racists like xyyman just simply favor their own people and hate other peoples.

The type of Racists you and Geneticker are is a mixture of the mental illness of German Dziebel with racism. People who suffer from it get a sadistic thrill from attacking random racial groups even if the attacks doesn't benefit their racial group. The more idiotic the logic of your racist ideas are the more you enjoy it. Anyone who confronts your ridiculous ideas receives vicious verbal insults.

I get the same disturbing fibe from your posts, German Dziebel's posts, and Geneticker's posts.

Davidski said...

@Rob

I'm pointing out that there's direct evidence of both males and females moving from the steppe to the Balkans during the Chalcolithic.

Whether you like it or not, this direct evidence contradicts your theory.

Rob said...

@ Dave

I'm merely pointing out that you lack knowledge
Vucedol I3499 has 'steppe admixture' - he is G2 and mtDNA T2 (both found already in Neolithic SEE)
And I never said some males no males moved from the steppe. Rather, what the paper clearly shows is that steppe & EHG admixuture long predates the date fixed to the Kurgan migration. There was bi-directional movement around the entire course of the Black Sea.
If you're deeply attached to the 'Dumb-Ass' view of European history, then you should try getting over it, maybe meditation ?

Rob said...

@ Sam

Very interesting. Kind of like nihilism ?
That's some profound analysis from a mouthbreather.

Gioiello said...

@ EastPole

And that -m- in *mems-ro- weren't interchangeable with -n- is demonstrated from Latin menstruum > *mens-ro- with the different outcomes: *msro- >*mbro and *nsro->nstro-.

Gioiello said...

If we take Latin monstrum certainly from moneo, we may hypothesize three different roots: *mem-, *men-, *mon-. To explain the link with moon, meter, meat etc, and why not Latin mentula etc

Samuel Andrews said...

@Rob,
"And I never said some males no males moved from the steppe."

Says mr. "I know more than everyone here put together"

@Rob,
"Very interesting. Kind of like nihilism ? "

Yes exactly. You nihilist are the most demented people on the planet. You, Geneticker, and German Dziebel exude the same cold blooded evil fibe in your posts. I prefer evil people who at least have some type of justification for being evil like Muslim terrorists or Nazis.

You should read the short story "Good Country People." You remind me of Joy-Hulga.

Arza said...

@ Rob
Vucedol I3499 has 'steppe admixture' - he is G2 and mtDNA T2 (both found already in Neolithic SEE)

There are three Vucedol samples:
I4175 3000-2700 BCE___ _______________________ F U4a ..
I3499 2884-2666 calBCE (4176±28 BP, BRAMS-1304) M T2e R1b1a1a2a2
I2792 2872-2617 calBCE (4130±35 BP, Poz-90126)__ M T2c2 G2a2a1a2a

I4175 [RRRRRROOOOOOOOOOOOOOOOGGGGGGGGGGGG]
I3499 [RRROOOOOOOOOOOGGGGGGGGGGGGGGGGGGGG]
I2792 [RRGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG]

Anatolian Neolithic (Grey), Yamnaya from Samara (Orange), EHG (Red)

Samuel Andrews said...

@Rob,
"Vucedol I3499 has 'steppe admixture' - he is G2 and mtDNA T2 "

Hey dumbass he had R1b Z2103 not G2.

Gioiello said...

@ Samuel Andrews

Should I thank you for not having put me amongst the Nazis like Rob, German Dziebel, Genetiker whom I consider some friends of mine, but, please, explain me why I should be against the Kurgan theory, being an R1b1a2-L23-Z2110 who should have come from there, and amongst the chiftains? I don't know your Y, but you do know mine.

Arza said...

@ Gioiello
why not Latin mentula

Those filthy Latins.

miętolić (mientolit') - crimp, squeeze something in his hands, knead, squeeze

Gioiello said...

@ Arza

A linguist studies words per se, and words non olent. That "mentula" may be linked with the other words is for the same membrum/mentula. Only Sicilians use the word now: mentula>minchia. In Italian we have only "minchione" ,because who gets a great mentula is thought stupid, but that is thought only from the other men, not from women.

Davidski said...

@Rob

Otherwise, you'd be aware that - unless specific data (known admixutre & migration Hx) is known- the GW variation rapidly becomes de-coupled from uniparental markers.

That's very funny, because I2181 is an outlier and clusters half way between Khvalynsk and Balkan Chalcolithic, so he's in all likelihood a first generation mixture.

Are you actually implying then that he inherited his Y-hg from his mother and mtDNA from his father.

If not, then what the fuck are you really saying?

Rob said...

@ Sam

Professor Freud can't even spell "Vibe"

Davidski said...

@Rob

You keep using the Vucedol samples as evidence to back up your arguments.

Please note that there are two Vucedol males, one with significant steppe admixture and the other with none. Care to guess which is which based on their uniparental markers?

Vucedol I3499 R1b1a1a2a2 T2e

Vucedol I2792 G2a2a1a2a T2c2


Btw, R1b1a1a2a2 is R1b-Z2103.

Arch Hades said...

So David whats ur problem with that paper from the Max Planck institute on Greeks, Albanians and Southern Italians? It seems to be arguing that mainland Greeks and Albanians are influenced by Migrations from Northeastern Europe (Slavs) While Cypriots, Island Greeks, Griko speakers in S. Italy..as well as Italian speaking Southern Italians are not...and are almost entirely descended from Neolithic and Bronze age migrations to Souheastern Europe.

Rob said...

@ Dave

I don;t know how else I can explain it. When populations have been in contact for thousands of years, the simple links you're attempting to make don;t hold, because it's not the situation as in 7000 BC when the earliest farmers were coming into contact with foragers, each with more or less neatly segregated autosomes correlating with uniparentals.

There are actually 3 Vucedol individuals, and the most east shifted is the female (U4a).

Davidski said...

@Rob

I don;t know how else I can explain it. When populations have been in contact for thousands of years, the simple links you're attempting to make don;t hold, because it's not the situation as in 7000 BC when the earliest farmers were coming into contact with foragers, each with more or less neatly segregated autosomes correlating with uniparentals.

Your argument doesn't hold in this instance, because the sample in question is an obvious outlier sitting half way between his parent group and the most likely admixture source.

Therefore, he's clearly a first generation mixture, and so his uniparental markers make perfect sense: R from the steppe and HV from the Balkans.

There are actually 3 Vucedol individuals, and the most east shifted is the female (U4a).

But no one gives a shit, because the two males show Y-haplogroups that match their genome-wide ancestry.

EastPole said...

@Alogo
“stop confusing language with genetics and assuming cognates have Slavic roots, man”

There is a correlation between genes and languages. It is a proven fact.
R1a-Z93 correlates with Indo-Iranian languages.
R1a-Z283 correlates with Balto-Slavic languages.

If some Indo-Iranian words, as I have shown, can be derived from Slavic words, then R1a- Z645 probably used similar words and they were close to Slavic. Because there are plenty of such words, we can conclude that CWC population in Poland was using language closer to Slavic than to Indo-Iranian.

Rob said...

Dave 25% is not an outlier
It is "normal" in a diverse and structured setting of the Balkans, which has had 10000 years of contact with the steppe
If we saw such a person in 3500 BC Britain, yes that's an "outlier"; not in SEE.

Davidski said...

@Rob

It is "normal" in a diverse and structured setting of the Balkans, which has had 10000 years of contact with the steppe.

If it's so normal then why aren't there any samples with such peculiar eastern-shifted heterogeneous genetic structure in the Mesolithic and Neolithic Balkan sets?

It's obviously a new phenomenon that started in the Chalcolithic. There are no grounds for you to claim that it has been happening for 10000 years.

Please stick to the facts instead of coming up with a creative new narrative to explain away the data.

Alberto said...

It's really sad to watch the kind of indigestion caused by this paper about SE Europe in some people. To be honest, this was a very mild dose of reality. I don't know what will happen when the real deal comes out.

From supp. Table 4.3, I'll post these numbers of qpAdm models:

Pop WHG AN Yamnaya
Balkans_Chaolcolithic_outlier 11.5% 40.2% 48.3%
Varna_outlier 14.6% 40.5% 44.9%
Trypillia_outlier 14.8% 50.4% 34.7%
Balkans_BronzeAge 11% 59.6% 29.4%
Vucedol 4% 79.3% 16.6%
Balkans_IronAge 0.2% 85.6% 14.2%

Notice that the Balkans_Chalcolithic_outlier is from 4500 BC and the Varna_outlier is from 4600 BC. They not only have 0% real Yamnaya admixture. They also have 0% EHG admixture. We don't have evidence of EHG in Ukraine until 3000 BC. These are samples with high SHG-admixture and some amount of CHG. And they plot around Bell Beakers from Germany (as the qpAdm models put them too).

These are outliers (like the Trypillia one), but from small sample sets. And what they mean is that this kind of ancestry was entering (irregularly) the Balkans from very early: first with local HGs (like is the Malak_Preslavets samples from 5600 BC) and then gradually increasing in CHG admixture, creating that pseudo-Yamnaya effect.

No one says that from 3000 BC some shepherds from the steppe didn't move to the Balkans. But then again, everyone was moving around and getting admixture from everywhere. David Reich spoke recently with probably quite more samples in mind than the ones we have here, so I would expect that the next paper will strengthen his statements, not turn them around.

Though seeing comments about the Mycenaean civilization being an offshoot of Sintashta/Srubnaya or even of CWC, it seems that some people have definitely lost their grip on reality.

Davidski said...

@Alberto

Look at the PCA and then try again.

The Chalcolithic and Bronze Age clines run from the Balkans to the Caspian Steppe, not to SHG or CHG. This has some important implications for Ukraine, but I won't get into that now.

CHG looks almost totally irrelevant to the Bronze Age Balkans (note the big white space). It might be more relevant in the Iron Age.

Do you know what that means? You probably do, but you'll prefer to ignore it. But please don't tell me how to read qpAdm output, because you're not in the position to do that.

Matt said...

Don't totally understand the discussion, as far as I can tell:

I2a2a1b is cosmopolitan between Globular Amphora Ukraine, Iron Gates HG, Latvia HG, Ukraine Neolithic and the two Yamnaya Bulgaria and Yamnaya Kalmykia samples. So the most likely centre of origin is really intermediate all of those, which may not be steppe (seems like it probably wouldn't be, but I haven't looked too closely at the map).

I2a2a1b1 is cosmopolitan between Ukraine Neolithic and the two Yamnaya Bulgaria and Yamnaya Kalmykia samples, then the Bronze Age Bulgarian samples which are older than the two Yamnaya (and older median dates than all the Yamnaya group). So most likely pre-Yamnaya Ukraine at the moment, (but more samples in future may contradict). Which probably means steppe, but seems atm likely no more eastern on the steppe than its westmost border in Ukraine.

Anyone have any thoughts on why I2a2a1b1 is extinct in the Balkans today and generally uncommon in Europe? Seems to have been gone by the late Bronze Age of the Balkans either also (y haplos G2, R1a, J2), when a richer and clearer signal of relatedness to Yamnaya proper shows up in the ADMIXTURE (supervised and unsupervised).

Folker said...

@Arch Hades

This paper is interesting if you keep in mind that they used ponly modern DNA, find in mondern populations. But, what is far less interesting, is their link with IE; They clearly tries to put a square in a circle. It's wishfull thinking, to try to save the "Anatolian Hypothesis". Mind who works at the Max Planck Institute. Somes are trying to save their butts.
The correct way to adress the problem it's to start with the facts, not the hypothesis.

Alberto said...

@Davidski

Look at the PCA and then try again.

The Chalcolithic and Bronze Age clines run from the Balkans to the Caspian Steppe, not to SHG or CHG.


Christ, I thought you could actually read a PCA with some slightly more complex mental processing.

So probably according to you, the Ukraine_Eneolithic sample is a mix of EHG and Anatolia Neolithic, right? How about you start considering 3 way admixtures?

This has some important implications for Ukraine, but I won't get into that now.

Can't wait to read about them.

Rob said...

@ Matt

You raise a good question about I2a2, which cannot be answered until we start getting middle Bronze Age - Iron Age samples from around Europe.

Rob said...

@ David

Yes. And why isn't the specific CHG/ EHG mix not found in the Mesolithic - Neolithic steppe either ?
In fact, why does it first appear in varna (4500 BC) before the Khvalynsk man (4000-3600 BC) ?
Why - you'll see in the next couple of papers. That's why
For now you can keep your creative writing up about I2a2.

Davidski said...

@Alberto

So probably according to you, the Ukraine_Eneolithic sample is a mix of EHG and Anatolia Neolithic, right?

There's no cline to suggest what it is. Could be a lot of things apart from the obvious Ukraine_Neolithic.

How about you start considering 3 way admixtures?

What makes you think I don't?

Clines still exist even in 3 way mixture models, they're just shifted away from the two main mixture sources.

How about you start paying attention to clines and quit the wishful thinking.

Forget the CHG/SHG fantasy. There's no evidence for it. You just made it up.

Davidski said...

@Rob

Yes. And why isn't the specific CHG/ EHG mix not found in the Mesolithic - Neolithic steppe either? In fact, why does it first appear in varna (4500 BC) before the Khvalynsk man (4000-3600 BC)?

It's already found somewhere on the Caspian Steppe at 4500 BC. That's why the Balkans Chalcolithic cline runs from the Neolithic Balkans to the least admixed Khvalynsk guys.

No idea how much CHG the Varna and Balkan Chalcolithic outliers have yet. Maybe very little. I need to check this myself.

Alberto said...

@Davidski

Forget the CHG/SHG fantasy. There's no evidence for it. You just made it up.

So there is no evidence for SHG-like presence in the Balkans (and Ukraine), from Mesolithic to Chalcolithic?

And CHG, we have Kumtepe6, 4600 BC, with CHG admixture. And Kumtepe is quite close to Varna, believe it or not.

In contrast, there is no evidence of EHG anywhere near Bulgaria (and not in Ukraine either) until 3000 BC.

But clearly you decided to start moving against the data, and not with it. So I'll leave it here.

Rob said...

@ Dave

Read Morgunova. Khvalynsk dates after 42/4000 BC.
Mathieson et al need to update the dating on their Supp Data File.

Olympus Mons said...

@Davidski .

STOP!!
Alberto says "And CHG, we have Kumtepe6, 4600 BC, with CHG admixture. And Kumtepe is quite close to Varna, believe it or not."

Now when I said KUM6 has my shulaveri because it had CHG admix, you flat out called me an Idiot. And I accepted. And said, ok, if KUM6 had no CHG admix than could not be my Shulaveri, and so I was being an Idiot.

So, is Alberto an Idiot as well?

Matt said...

D-stats in the paper indicate that the shift between the "Balkans Bronze Age" compared to Chalcolithic is strongest to EHG, then to Yamnaya, then WHG, then CHG last. I don't see a lot of evidence there at least, that the Balkans Bronze Age has a particular thrust of CHG related ancestry.

Though this I think throws together what look like at least 4 clusters of individuals with different ADMIXTURE profiles, who might be different in ancestry.

Re: PCA generally, it would be a lot easier to tell about CHG with a dimension 3 that is usually more of a specialised CHG related dimensions.

Davidski: It's already found somewhere on the Caspian Steppe at 4500 BC. That's why the Balkans Chalcolithic cline runs from the Neolithic Balkans to the least admixed Khvalynsk guys.

When I pool together the samples labelled Balkans_Neolithic, Balkans_Chalcolithic, and Trypillia, they don't really seem to go the the Eneolithic Steppe at all:

http://imgur.com/a/FzHuo

Seems like it goes to Ukraine Neolithic (which is surprising because the Iron Gates HG should be closer, but not as surprising as anything involving the Eneolithic steppe).

But it's quite difficult to call, and I can't say Alberto is wrong that the cline for Balkans Neolithic->Chalcolithic actually points to SHG / the most Western Ukraine Neolithic (basically the same thing).

If we're talking about fitting the Balkans Chalcolithic outliers as about 2/3 Balkans_Chalcolithic, 1/3 Samara Eneolithic though, that makes a lot more sense than fitting with the Yamnaya, which doesn't really seem to work very well.

I think compared to if you modeled the Balkans Chalcolithic outliers using the Iberian cline (Anatolia->Iberia_EN->Iberia MN) or Central European cline (Anatolia->LBK->Central_MN->Globular Amphora) then using a Balkans Chalcolithic +Steppe Eneolithic model instead substantially cuts the Steppe_Eneolithic ancestry. But I'm really unconvinced you could use a Ukraine_Neolithic+Anatolia_Neolithic+CHG model or anything like it to explain the Balkans samples parsimoniously.

Alberto said...

@Matt

Thanks.

I'm really unconvinced you could use a Ukraine_Neolithic+Anatolia_Neolithic+CHG model or anything like it to explain the Balkans samples parsimoniously.

Looking at the 3 Varna samples that seem to have some amount of CHG admixture (maybe some 10%?), then the difference between the 3 would be the amount of SHG-like admixture (or Ukraine_Neolithic-like, if you prefer, for geographical reasons). Whether that's parsimonious or not might be determined by the alternatives. At 4500 BC in Bulgaria, do you think there is an alternative model? (the paper proposes modeling the outlier as 45% Yamnaya, for whatever reason).

Davidski said...

@Matt

I don't know what you're trying to demonstrate with those plots exactly, but obviously contacts between Khvalynsk and the Balkans, and specifically Varna, happened during the Chalcolithic.

There may have been some contacts between the Balkans and Ukraine during the Neolithic, but that's obviously irrelevant to where people like Varna outlier got the steppe admixture.

In other words, basing clines on irrelevant samples is probably not a good idea.

Matt said...

Well, I'm just trying to respond to your comment really. You talked about a cline from the Chalcolithic Balkans to Khavalynsk. I see no such cline if you're using the terms as they describe. That is all I'm doing by showing the clines, as I didn't want the erroneous idea that there was a Balkans Chalcolithic cline leading towards Khavalynsk to get repeated by a game of telephone.

The outliers do fit as most likely to be similar to the Chalcolithic Balkans plus Khavalynsk, as far as it looks.

You could probably estimate the outliers as on a HG-heavy position on the actual Balkans Chalcolithic cline, beyond what is actually present, plus some CHG, but I think there's no actual reason why you should, and it would be complicated to do so. I'm not explaining them that way.

So what's the significant disagreement?

Grey said...

Romulus said...
"According to the paper Varna and Trypolie have 10% EHG but no CHG."

I think people defining "steppe ancestry" as EHG + CHG obscures more than it reveals as unless i'm missing something there was clearly a time before they mixed when EHG was "steppe ancestry."

If correct then there's two steppe ancestries:
- steppe ancestry A (EHG)
- steppe ancestry B (EHG+CHG)

and the question that follows is did steppe A move into WHG (or CHG) territory before PIE or was there an early PIE (just EHG) and a later PIE (EHG+CHG).

Grey said...

Ric Hern said...
"I see everybody forgot about Sredny Stog. When we look at the spread of animal domestication it clearly ends up first in Sredny Stog before Khvalynsk and Yamnaya."

Just to repeat there's a later model (Ertobolle culture) of sedentary, pottery using wetlands HGs adopting animal husbandry from adjacent farmers.

(the sedentary and pottery using are significant as signals of food source / population density)

So if sedentary, pottery using HGs extended around the northern and western shores of the Black Sea then maybe it was them?

#

"It is also proposed that Yamnaya formed due to a Sredny Stog Migration into Khvalynsk."

If coastal Black Sea HGs on the western shore did adopt animal husbandry then it might have spread to the north shore and from there (minus the wetland HG part) to the steppe itself.

Grey said...

Matt
"Anyone have any thoughts on why I2a2a1b1 is extinct in the Balkans today and generally uncommon in Europe? Seems to have been gone by the late Bronze Age of the Balkans either also (y haplos G2, R1a, J2), when a richer and clearer signal of relatedness to Yamnaya proper shows up in the ADMIXTURE (supervised and unsupervised)."

My thoughts are...

1) ydna I (various clades) was originally paleo WHG across most of Europe, including around the shores of the Black Sea but they were mostly replaced by the farmers except in a few protected regions - mostly mountain/swampy terrain the farmers couldn't use.

2) some sedentary ydna I HGs in the coastal areas around the Black Sea (whether mixed with ANF or not) adopted animal husbandry and became farmer/HGs (like Ertobolle) with the ones on the western shore having more farmer contact and the ones on the north shore having more steppe contact.

3) both later spread back into Europe as a result of PIE - firstly the more farmer-ish ones getting pushed away from the steppe by PIE (Hungary ydna I farmers? Sardinian interior?) and secondly the ones who had become attached to Yamnaya (Remedello?).

(aka at least part of the WHG resurgence)

nb the key part of this is my assumption there were three distinct biomes involved:
- hill country away from the Black Sea coast to the south and west
- steppe to the north
- coastal wetlands around Black Sea coast adjacent to and connecting the other two regions

also maybe a similar pattern with different components around the Caucasus and Caspian.

Matt said...

@ Alberto: At 4500 BC in Bulgaria, do you think there is an alternative model? Probably using pre-Yamnaya steppe cultures like Khavalynsk, but not actually a population as heavily CHG as Yamnaya.

Various clines are possible here though, purely on the basis of PC1 vs PC2. For instance why you could even model the Balkan Bronze Age as varying mixtures of the Chalcolithic outliers and mainstream, plus the Anatolian Bronze Age - http://imgur.com/a/vRKA2.

@Grey: steppe ancestry A (EHG) Or were Steppe Ancestry A speakers of some Uralic like language, and only took on a late PIE dialect?

The linguistic reconstructions of PIE that favour the steppe hypothesis "Neither woven wool textiles nor wheeled vehicles existed before about 4000 BCE. It is possible that neither existed before about 3500 BCE. Yet Proto-Indo-European speakers spoke regularly about wheeled vehicles and some sort of wool textile. This vocabulary suggests that Proto-Indo-European was spoken after 4000–3500 BCE". Khavalynsk Culture probably just about within the time frame and material culture, and this culture having some sporadic influence on the Balkan Chalcolithic.

As I understand it, hard to say anything about cultures earlier than that (pre-Khavalynsk), other than that if they spoke a language meaningfully ancestral to PIE, we can't really say so certain (because that seems not to be the stage at which any language expansion happened).

mostly replaced by the farmers except in a few protected regions - mostly mountain/swampy terrain the farmers couldn't use.

I think you might be being a bit optimistic about how much of the European territory that the early Neolithic groups (Cardial, LBK) could actually use effectively!

Also I don't know if I understand you correctly, but is part of your idea that the WHG resurgence is actually from a Balkan migration of a single population? I don't know if that is supported by Lipson, Olalde etc which support structure among Neolithic populations and absorption of local HG groups.

Grey said...

Matt

"Or were Steppe Ancestry A speakers of some Uralic like language, and only took on a late PIE dialect?"

Yes, I need to be more careful with labels - when I say PIE i'm generally thinking of a semi-nomadic steppe herding culture rather than the language. So what i'm really saying is if steppe ancestry B (EHG+CHG) is from Yamnaya-like herders then is steppe ancestry A (EHG alone) from HGs or from early herders before the CHG mixture (or a bit of both).

#

"I think you might be being a bit optimistic about how much of the European territory that the early Neolithic groups (Cardial, LBK) could actually use effectively!"

Yeah, I can see why it would look like that but I agree they had a very wide but shallow distribution and this is significant for the scale of their disappearance.

What I'm thinking here is relative population density. Assuming for the sake of argument a 10:1 ratio of population density between farming on suitable terrain and average HG then even if the farmers only took 20% of the land the population ratio would end up being 5:2.

30% would be 30:7 ~ 4:1
40% would be 40:6 ~ 7:1
50% would be 50:5 ~ 10:1

so i'm thinking the HG population would be dwarfed by the farmer population even with the same amount of land.

#

"Also I don't know if I understand you correctly, but is part of your idea that the WHG resurgence is actually from a Balkan migration of a single population?"

No - i think it may have happened multiple ways but one part may have been WHG-like farmers from west of the Black Sea.

Other examples might be WHG groups around the northern Black Sea who had become associated with Yamnaya and spread with them or maybe Ertobolle-like groups around the Baltic/North Sea etc.

Basically I used to think the WHG resurgence was an "aboriginal stockmen" model where HGs from the mountains became incorporated into farmer groups as workers/herders. Now i wonder if it was more that some WHG groups from specific regions adopted either a farmer or herder culture and then spread with those same farmers/herders.

Alberto said...

@Matt

Yes, many clines are possible in 2 dimensions and with few and diverse samples.

At 4500 BC in Bulgaria, do you think there is an alternative model?

Probably using pre-Yamnaya steppe cultures like Khavalynsk, but not actually a population as heavily CHG as Yamnaya.

Yes, some pre-Khvalynsk (Samara culture?) moving to Bulgaria, even if no traces in Ukraine. Since Khvalysnk/Samara Culture derive from EHG, then if you draw a line leading to EHG, the Varna_outlier deviates very little to the east, so it probably will only require 2-4% CHG. If instead you draw a line to Ukraine_Neolithic, that would require quite a bit more CHG, maybe 10-15%. So I hope it will be possible to check when we get the samples.

Matt said...

@ Alberto, hmmm... I think like Balkans_Chalcolithic+Ukraine_N+CHG sounds like a bad fit for the outlier compared to Balkans_Chalcolithic+Samara_Eneolithic as it's more complicated.

Though you mention Samara Culture, I think the outliers could also work as Balkans_Chalcolithic+Ukraine_Eneolithic, and Ukraine_Eneolithic may be able to fit as Ukraine_Neolithic+CHG+Anatolia_Neolithic as they seem to suggest with their D-stats. (Ukraine_Neolithic itself seems to fit as Ukraine_Mesolithic+WHG.) If the one Ukraine_Eneolithic sample is typical, and as you say we don't know.

That would still be considered steppe movement (as the steppe IRC extended into these areas of Ukraine), but not one involving populations quite like those from Samara / Kalmykia. It may also make more historical and archaeological sense? As IRC the archaeology is specific that the Chalcolithic Balkans is drawing. But I'm no expert on the archaeology.

However, the Western Ukrainian Yamnaya later in history (500 years?) are quite like the Samara group...

....

On another note, bit off topic, but I've been just been looking at the studies and thinking about what Mathieson did not include in their PCA.
Off the top of my head, none of the Mittnik 2017 samples were included, so no Swedish TRB Funnel Beaker, no Kunda / Narva (looks like they've sampled different Latvian HG), no Baltic Bronze Age, no Pitted Ware or Combed Ware. None of Lipson 2017's, Hungarian Late Copper Age, Blatterhohle German samples either. Olalde 2017's sample set of Beakers and Middle Neolithic Atlantic Europe is also not included.

So the huge PCA in this paper, does not cover all the samples in depth we have for ancient Europe+West Asia so far.

Just comparing the sample IDs, it looks like there may be up to 919 unique IDs between the papers, and even this huge paper only covers 470 and Olalde 481 (though this might overestimate some of the same samples where they've slightly altered the IDs). There's almost certainly the Iron Age Scythian samples not used as well (and also Fu et al Ice Age, though I think those samples are quite specialized and probably few have real relevance to later periods, though they may help firm up regional WHG differences, and I guess the British Isles Iron Age series).

Arza said...

@ EastPole

I've managed to overlay SE Europe PCA onto N Europe PCA. These lines start to look really good.

https://2.bp.blogspot.com/-5c96SyPJZ_M/WSg3X-W4sGI/AAAAAAAAAL8/NtsN3YerdacyTZ8OTJqCt8NvL0WmA4z4gCLcB/s1600/Baltic_SE_Europe_PCA.png

Gardeners said...

Why do people assume early farmers were men? Is there any evidence other than the assumption? I would suggest farming was invented by women as a takeover from the gathering part of hunting and gathering. We know that from Papua and Aboriginal Australia that women were re-planting myrniong, a tuber with a yellow flower, in order to get another crop. They were also doing the harvesting of seeds to make unleavened bread, and broadcasting that seed for the next season's crop. The men were off chasing game. Hunting men would have considered farming beneath them-I know that is an assumption on my part.

chris said...

I am Greek and my haplogroup is a subclade of r1a-z93. I found out that I share common segments with this sample on gedmtch. I don't think he is Myceneann because when Mycenenans arrived to Balkans they had mixed with neolithic populations and their autosomal results should be quite different. Proto-Thracians on the other hand is suggested that they migrated from pontic steppes (corded ware culture) during Bronze Age, they mixed with local neolithics Balkanic populations and during the Iron Age they evolved to Thracians and Dacians. Corde Ware culture had interactions with Srubna. What I am thinking is that among the proto-Thracian people who left the steppes there were some individuals with origin from Srubna culture (r1a-z93). So this sample could be a Proto-Tharcian. What do you think?

Davidski said...

@Chris

Mycenaean samples have steppe ancestry very similar to that of the R1a-Z93 sample that you're talking about. See here...

Steppe admixture in Mycenaeans, lots of Caucasus admixture already in Minoans (Lazaridis et al. 2017)

But ancient steppe ancestry is generally very similar, because most of the Bronze Age steppe groups were closely related. So it's still up for debate who the Mycenaeans got their steppe ancestry from exactly. See here...

Main candidates for the precursors of the proto-Greeks in the ancient DNA record to date

chris said...

Lazaridis paper suggests a steppe admixture on Mycenenans. But Mycenenans also brought J2a arriving in Greece so I assume they had already mixed with neolithic populations outside Greece and the migrants who arrived to Greece were autosomally a mix of IE and pre-IE people and as a result different autosomally from the ancient Bulgaria r-z93 sample. Am I right?

Davidski said...

The R1a-Z93 sample from Bulgaria is actually a very good fit for the genome-wide (autosomal) steppe ancestry in Mycenaeans.

And so far there's just one Y-chromosome haplogroup result from Mycenaean Greece, so I wouldn't get too hung up on that. Obviously, you can't make a judgment about the Mycenaean Y-chromosome gene pool based on one sample.

chris said...

@Davidski
What I find starnge is that in modern Greece and Balkans R-Z93 is very rare. If Proto-Greeks and/or Thracians had R-Z93 why it is so uncommon nowadays?

Davidski said...

@Chris

If Proto-Greeks and/or Thracians had R-Z93 why it is so uncommon nowadays?

Maybe because they lived in the Balkans ~2,500 years ago, and a lot has changed there since then, with the elites being replaced several times over?

chris said...

Yes, you have a point but we can assume R-Z93 was not frequent even in ancient times. I say so because haplogroups R1b,E-V13 and J2 are quite frequent in Greece and South Balkans and exist in area since Neolithic or Bronze Age. In addition south Italy has these ancient Greek haplogroups but R-Z93 is very uncommon in this place too. What I think is possible is that Indo-European settlers of Balkans derived more from Corded Ware culture but due to interactions with Srubna culture a limited number of these settlers were R1a-Z93 from this Srubna.

«Oldest ‹Older   201 – 281 of 281   Newer› Newest»