Surplus EEF ancestry in modern day Slavs

I think that most people interested in archaeogenetics would easily name two ancient groups that are the base of genetic composition of modern day Balto-Slavs. These two groups are the Baltic Bronze Age and the Corded Ware Culture.

But nor the Slavs, nor even the Balts lie on a hypothetical cline between CWC and Baltic_BA. They also aren't identical to any of them. Clearly at least one additional source of ancestry is needed - the neolithic farmers of Europe.

To check which neolithic group can or cannot be the source of this surplus ancestry I've modelled modern populations using a large set of ancients. Then I subtracted from the targets the ancestry from sources not located on the EEF-WHG cline.

Below you can see the West Eurasian PCA with the distribution of samples used in this model.


 

Results of this experiment:

Slavs


Balts


Germanics


Selected Germanic, Baltic and Slavic groups


Selected IA and Medieval groups


The first thing you may notice is the stark difference in the type of the EEF ancestry (that hasn't been mediated via CWC or Baltic_BA) between Germanic and Balto-Slavic populations. 

The second, even more important thing, is how "southern" the surplus EEF ancestry in Slavs is.

The third one - that the Balts share it with the Slavs.


Consequences of this are pretty obvious. It's impossible to explain the alleged "westernisation" of migrating Slavs by the assimilation of some random Germanic tribes on their way to their new homelands as the type of the EEF ancestry doesn't match.

This alleged "westernization" cannot be explained by the assimilation of any local group at all, as the same ancestry is shared with the Slavs who supposedly have stayed in the Slavic homeland or even with the Balts.

It's also unlikely that this ancestry came from the Baltics, Belarus or even Ukraine or Poland. LBK has been biting the dust for a long time, other groups had too much WHG and Ukraine Neolithic outlier was an, well, outlier. Most likely from somewhere in the Balkans.

Selected mesolithic to eneolithic groups from Ukraine

 

Another method to show the difference in the type of the EEF ancestry is to create a three-dimensional reprocessed PCA as seen below. Notice that the plot was rotated in a way that Baltic_BA overlaps with Yamnaya. This nullifies the difference between them and allows us to look at the non-Steppe, non-Baltic_BA ancestry of the populations of our interest.

This method also deals with the misguided argument that using Baltic_BA in the models artificially forces the southern ancestry onto Balts and Slavs because it clearly shows that even if the Baltic_BA population would be identical to the Yamnaya cluster or non-existant in Balto-Slavs, the southern ancestry would still be the preferred one.

Reprocessed 3D PCA


Sidenotes:

Some Slavic and Baltic groups show a shift towards the GAC-like EEF ancestry. This may be due to the assimilation of Germanic populations in the historical times, but also due to the assimilation of earlier local post-CWC groups with higher GAC and/or WHG ancestry.

Some Slavic groups (e.g. Ukrainians) show slightly elevated CHG ancestry. Source of it is unclear (assimilated steppe groups? leftovers from Greek colonies in Ukraine? backflow from the Southern Balkans related to christianisation?).

Some samples from Pohansko or Krakauer Berg show EEF-profile close to the Germanic one. This may indicate that they were mixed with Germanics and that their non-Germanic ancestry was of the Baltic or even Baltic_BA type. It's a clear signal of heterogeneity in these populations and a possible indication of contacts with the North.

Out of the three samples from Libiva the Germanic one (LIB2) is an outlier, while Italian-like LIB3 and Slavic-like LIB11 show the same "southern" EEF-profile compatible with the one seen in Balto-Slavs.

Comments

Angantyr said…
Interesting. I think it would be very interesting to see where the Tollensee warriors plot.
Arza said…
@ Angantyr

Tollense has additional WHG and for some samples even Wartberg isn't enough and they pick Iron Gates.

https://i.postimg.cc/rMRzjZmz/Tollense-EEF-WHG.png

Note: WEZ56 in other dimensions is shifted from others and it seems that he has more of the Baltic_BA-type of WHG than even the Baltic_BA itself.
ambron said…
Arza, colleagues at AG talked a lot about this entry. They even ask you some questions. I don't know if you are following this? I don't understand, why don't they come here to talk?
Arza said…
I'll read this discussion later, but I rather won't reply before Sunday.
ambron said…
Arza, what did you mean by writing this:

"Some samples from Pohansko or Krakauer Berg show EEF-profile close to the Germanic one. This may indicate that they were mixed with Germanics and that their non-Germanic ancestry was of the Baltic or even Baltic_BA type. It's a clear signal of heterogeneity in these populations and a possible indication of contacts with the North".

Probably not the Pripyat basin. In the Middle Ages, "pure" Baltic BA could only be found probably on Gotland...
Arza said…
Both Krakauer Berg and Pohansko seem to have more complicated histories than that some Belarusians came there and mixed with local Germanics, but for now don't read too much into this. I just wanted to indicate what's worth checking in the future, because a dedicated model is definitely needed here.

And yeah, I'm thinking about the kind of samples that we have seen in Margaryan et al.
Arza said…
I think that the next post will be about the Krakauer Berg. I've managed to separate them into two nice clusters. I'll fine-tune the model and redo everything and if it give the same result I'll post it.
ambron said…
This is great, because Krakauer Berg is also a current topic on AG.

Arza, where can we find Poles on this cline from the last PCA? As I understand it, Poles should be located somewhere behind the Belarusians in the direction of DEU LBK. Can you add Poles?
Arza said…
https://i.postimg.cc/c1xqzQr5/Vahaduo-Custom-PCA-7.png
ambron said…
Thanks! Cool! Now it is clearly visible that Poles cannot be a mixture of Belarusians and Restgermans. Only individual Poles can be perceived as having a Germanic admixture. But also in some Scandinavians see the Slavic admixture.
Arza said…
The one in Swedes is rather not Slavic. It looks mostly Baltic and/or Uralic.
The one in Poles - at least partially Germanic. But it may be also an elevated CWC/WHG from another source.
Erik Andersson said…
Hello Arza.
About the 3D PCA: The WHG in Norwegians is not accounted for by Yamnaya Samara, but the WHG in Belarusians is by LVA BA, leaving only EEF at the other end. Here is a simplified version of your PCA I put together, with similar result: https://i.postimg.cc/054FVmTN/LVA-BA-PCA.png
I created a ghost based on LVA BA with 30% of its HG ancestry subtracted and used that instead: https://i.postimg.cc/QxWPwCDJ/Ghost-PCA.png
The result is different, because now the WHG in Belarusians is not accounted for by LVA BA. So I don't think that this method proves that the EEF in Balto-Slavs was mediated by a source poor in HG ancestry. The only reason why it appeared that way is because the Balto-Slavic cline was defined by LVA BA, which has a lot of WHG. If I'm wrong, please explain why.
Arza said…
@ Erik Andersson

Hi!

If we assume that the WHG in Balto-Slavs came from a WHG-rich farmers, then Balto-Slavs can no longer have the Baltic_BA ancestry (because you'd be counting WHG ancestry twice).

Such scenario of course would be possible if not the fact, that PCA "sees" the difference in the type of WHG ancestry in Baltic_BA and other populations (especially GAC) and thanks to this we can be sure that it was mediated via Baltic_BA.

Besides that if you create a ghost you need to check if the outcome makes sense (e.g. that it follows the known clines, it's on an intersection of clines and so on).

Below is a West Eurasian 3D PCA (PC1, PC2, PC6) of the LTU_BA-30%LTU_Narva ghost:
https://i.postimg.cc/52QLM7P4/Vahaduo-Eurogenes-G25-West-Eurasia-3-D-5.png
https://i.postimg.cc/jdFPm043/Vahaduo-Eurogenes-G25-West-Eurasia-3-D-6.png
While mathematically correct, such source of non-WHG Baltic_BA ancestry would mean that Baltic_BA had no steppe ancestry at all.
Erik Andersson said…
@Arza
Thank you for the reply.
Yes, this is assuming that Balto-Slavs do not descend directly from Baltic BA, or a similarly WHG-rich and EEF-poor population.
I don't see why this would mean that Baltic BA would not have steppe ancestry. The ghost clusters with steppe-rich individuals in other dimensions.
https://i.postimg.cc/ZRY1kmgL/Vahaduo-Eurogenes-G25-West-Eurasia-3-D.png
It just has a lot of Balto-Slavic drift since it's based on LVA BA. LVA BA has a lot of steppe ancestry, so subtracting the WHG from it will leave you with a ghost with even more steppe ancestry, but also a massive amount of drift.
Arza said…
@ Erik Andersson

See the second plot I posted. There is a gigantic mismatch in the sixth dimension.

The distance to the nearest CWC sample is ~0.11. That's half the distance between Yamnaya and Kotias. Quite a lot!

And even if the distance was smaller, you still couldn't ignore this drift because it can't be recent. Samples much older than Baltic_BA show it too.
ambron said…
Arza, and what will Kowalewko look like at WE PCA?
Arza said…
Combined files plot on top of Lithuanians, close to the border between Lithuanian and Swedish clusters. This indicates that there is a lot of WHG in them.
ambron said…
Have you tried to throw individual samples of Kowalewko to WE PCA?
Arza said…
Yes, the outcome is generally similar to G25.
ambron said…
Will it be a hassle for you to provide a link to such a PCA? Because I have the impression that Kowalewko can group on it with the top Tollense cluster.
Arza said…
Just don't take this PCA too seriously, especially in the case of Kow24.

https://i.postimg.cc/k7HpJM8X/Vahaduo-Eurogenes-WE-3-D-1.png

They are like Scandinavians/post-CWC with at least some Balto-Slavic admixture, as evidenced in the G25, so completely different than Tollense.
ambron said…
Arza, thanks! I wanted to see it, because some time ago we had a leak that Kowalewko would be similar to today's Balts. So I suspect the informant saw the position of some samples on a PCA similar to the WE PCA.
Arza said…
And that's also the reason why Figlerowicz thinks that the common people didn't change, only the elites. He sees that the Y-DNA in Wielbark is different than in medieval samples, but on a classical West Eurasian plot he sees that autosomally Wielbark and medieval samples are nearly identical. The problem is that most likely they aren't.
ambron said…
It just needs to be remembered that when someone wants to show the weak genetic links of Tollense with the Balto-Slavs and Poles, he shows Wez samples at WE PCA.

However, this works both ways, because at WE PCA most of the Kow samples are grouped with Balto-Slavs and Poles.
ambron said…
Arza, how is the Krakauer Berg case? Progressing...?
Arza said…
Sloooowly.
ambron said…
Arza, do you have any news when we can see something new in genetics on the issues of interest to us?
Arza said…
Nothing. There are interesting presentation scheduled on ISBA9 but not even the abstracts are available.
ambron said…
They wrote:
"The current programme is only indicative. It will be confirmed following abstract selection".
So we have to be vigilant!
Arza said…
I'm afraid that it was written before the pandemic, rescheduling the conference to 2021 and turning it into an on-line event.

It's long after the "abstract selection".
ambron said…
Parastais on AG asks:

Maybe Arza could check if “Scythians” BCE in Ukraine already had this kind of EEF needed for Balts and Slavs.
Arza said…
Ukraine, maybe besides areas bordering the Carpathians, rather didn't have such ancestry.

There are 3 Scythian samples from Ukraine that fall more or less in the range of modern day Slavs and Slavic-related populations, but counting them as a source of this "Neolithic" shift is problematic, to say the least.

For example it'd mean that Slovaks, who genetically and linguistically seem to have central position among the Slavs, have zero Proto-Slavic ancestry.

What's even more important is what kind of ancestry Ukraine certainly did have. Ukraine was flooded with a variety of Asian admixtures, which lasted till the advent of the Chernyakhiv culture (Jarve 2019).

https://i.postimg.cc/wqtSCtXL/Scythians.png

Even from Moldova there are samples with a clear Asian admixture. If someone wants to see Central or Eastern Ukraine as a source of an even part of the Proto-Slavic gene-pool then we need to ask - where is the Asian admixture among the early Slavs? Where is the R1a-Z93 and uniparentals typical for Asian populations?

Moreover, as you can see on this plot, Slavic-like Scythians from Ukraine are likely migrants from the west, as they overlap with the main cluster of Hungarian Scythians. As a cherry on top - scy009, who is a Ukrainian Scythian that is the most similar to any of the modern Slavic samples has typically West-Central European Y-DNA R-P312 and from the ancient samples it's closest to a Middle Bronze Age specimen from Hungary (HUN_Fuzesabony_MBA:I20772).
ambron said…
Arza, the intention of Parastais' question was whether the Ukrainian Scythians could be the source of the westernization of the Balts and the alleged Pripyat Slavs. From our earlier discussion:

parastais:

Yeap, and Scythian_UKR shows that such genes were already in Ukraine. Every ingredient needed to produce Av2 was in Ukraine, so no need for some eternal Poles to come into mix..

ambron:

If the early Slavs turn out to be rich in P312, we'll talk. At the moment, he looks like a delegate from the Scythian ChotyƄec, as he looks autosomal like a inhabitant of today southern Poland.
ambron said…
So we see the case of the Ukrainian Scythians similarly.
Arza said…
I see that my reply was too nuanced. I have to use much simpler language, lol.
Arza said…
@ ambron

Don't copy-paste any of my comments and don't paste any links to this blog on Anthrogenica. Reason is below.

"Most of the argumentation left for the Central European homeland here is now based on circular and ultra-nationalist reasoning" - moderator of Anthrogenica.

BTW, isn't it strange that when this user, who poses as Indo-Aryan/Balto-Slavic and constantly trolls in Slavic-related topics, posted the plots, the screenshots were named "Unbenannt.png"? And isn't it even stranger, that he quickly and silently changed hosting of these files from imgbb to imgur, which doesn't reveal the filenames?