Southern Arc or Southern farce?—2.0
A "Southern Arc or Southern farce?" follow up
Before we begin with the technicalities, I apologize for the title. That’s the best I could come up with (LOL). Anyway, so I showed the signs for Anatolian/Levantine-related ancestry in samples from Eneolithic piedmont steppe in my previous substack post “Southern Arc or Southern farce?”. So, what’s new in this post? In the current post, the main highlight will be the working qpAdm models for Eneolithic piedmont steppe samples (something I couldn’t showcase in the previous post). I’ll keep this post short since there’s not a lot to uncover but just an addition to what I’ve already discovered. One thing that has changed is that I no longer primarily use ADMIXTOOLS for these tests since it’s too slow. We’ll make use of ADMIXTOOLS 2 which is significantly faster than the previous software package. There’s not much difference between the two with respect to results, so let’s give it a shot.
Relabelling
After running some informal tests on G25, I found that the sample PG2001 (from Progress-2) is closer to VJ1001 (from Vonyuchka-1) than to PG2004. So, we should label them accordingly.
Russia_Steppe_En_1 - PG2001 & VJ1001
Russia_Steppe_En_2 - PG2004F4-statistics
Z-scores for f4(Mbuti.DG, Turkey_N, CHG, Russia_Steppe_En_1), f4(Mbuti.DG, Levant_PPN, CHG, Russia_Steppe_En_1) and f4(Mbuti.DG, Levant_PPN, CHG, Russia_Steppe_En_2) are significantly positive. This indicates that these Steppe_Eneolithic populations have significant affinity towards Anatolian/Levantine-related populations relative to CHG. This result is not different from the f-4 stats results in my previous post. From this, I can predict that Russia_Steppe_En_1 might have more of this Anatolian/Levantine-related ancestry.
Admixture modelling
In the previous post, I had predicted that Steppe_Eneolithic has about ~15-30% of this “Southern Arc” ancestry and it was likely from Darkveti-Meshoko culture. And I was right in predicting so…
Russia_Steppe_En_1
Russia_Steppe_En_2
These are the only working models I could find for Steppe_Eneolithic samples. The I0434 sample from Khvalynsk culture can be modelled as ~42% EHG, 37% CHG and 21% WSHG (Lazaridis et al., 2022 had posited Siberian ancestry in Khvalynsk samples). You can find the qpAdm output in supplementary files provided at the end of the article.
Implications
Now, we summarize all of my findings-
EHG-related and CHG-related populations admix in the second half of the 6th millennium BCE to form populations at Khvalynsk and Piedmont steppe.
Steppe_Eneolithic samples definitely have Anatolian/Levantine-related ancestry.
Yamnaya has WHG-related ancestry.
Yamnaya get modelled successfully only when Steppe_Eneolithic-related, EEF and Ukraine_N-related populations are the sources in a rotating outgroups qpAdm model.
Therefore, Yamnaya most likely get this “Southern Arc” ancestry indirectly from Steppe_Eneolithic-related population itself rather than directly as shown in Lazaridis et al., 2022’s admixture modelling.
This surely messes up Lazaridis et al., 2022’s proposed timeline for the arrival of “PIE” (Core PIE) languages in north of Caucasus from Southern Arc (only if the Proto-Indo-Anatolian or archaic/early PIE homeland is in Southern Arc).
At last, I will also express my anguish for using such unconventional and confusing terminology. Better terms could’ve been used in their paper. And this is will be my last post on this topic.