Thanks to Skourtanioti et al , 2020, We now have samples that can help to fill in the picture of how populations in West Asia formed. Some of the most important, in my opinion, are the Halaf-era samples from Tell Kurdu and the Shulaveri-Shomu individuals from Azerbaijan. These were a couple of the important groups yet to be seen.
One of the interesting individuals is ART038, a Chalcolithic individual from Arslantepe. The interesting part of this individual is that they are R1b-V1636. While the paper does not delve into this individual much or look at complex models, I wanted to double check on this myself.
First things was to make sure that it was Halaf (Tell Kurdu) and not another group like Buyukkaya_EC that delivered the Anatolian ancestry. Several different looks at this in qpAdm confirmed for me that the Chalcolithic of Eastern Anatolia and the Caucasus required excess Levant PPN ancestry found in the Tell Kurdu individuals.
To flesh out any Piedmont ancestry and the different streams such as CHG and Iranian, I used the Late Neolithic individual from Ganj Dareh in the left pops (pleft) and Meshoko, with Ganj Dareh in the right pops, or outgroups (pright). The populations in the pright are Chimp, Ust_Ishim, Kostenki14, Brazil_LapaDoSanto_9600BP, Yana, WHG, EHG, Tianyuan, Mongolia_N_North, Taiwan_Hanben, West_Siberia_N, Levant_N, Ganj_Dareh_N, IBM, Barcin_N and Meshoko. This group allowed for the best resulting outputs.
Not only did I look to see if EHG or Piedmont ancestry is in these Chalcolithic samples, but even in the Late Neolithic Shulaveri-Shomu. Interestingly enough, they can be modeled with minor ancestry from a group like the Piedmont Eneolithic.
Shulaveri-Shomu
left pops:
Shulaveri_Shomu
TellKurdu_EC
CHG
Iran_LN
Piedmont_Eneolithic
numsnps used: 286501
best coefficients: 0.617 0.143 0.152 0.087
std. errors: 0.026 0.077 0.047 0.032
fixed pat wt dof chisq tail prob
0000 0 12 6.866 0.866351 0.617 0.143 0.152 0.087
While not completely necessary, Shulaveri Shomu can be modeled as deriving ~9% of their ancestry from the Piedmont.
Armenia Chalcolithic
left pops:
Armenia_ChL
TellKurdu_EC
CHG
Piedmont_Eneolithic
numsnps used: 471814
best coefficients: 0.483 0.226 0.291
std. errors: 0.016 0.029 0.023
fixed pat wt dof chisq tail prob
000 0 13 20.801 0.0769555 0.483 0.226 0.291
Armenia is a bit of an enigma here. It does stand out from the Shulaveri Shomu and Hajji Firuz individuals in that they lack any Iran_LN ancestry. The ancestry there is never larger than the standard error, which was around 6%. It seems that we may not have a great representative for Shulaveri Shomu in Armenia. I also looked to see if adding an Iranian Plateau or Central Asian group would help, but they were small contributions and never above the standard error. So, those models were left out.
Arslantepe Late Chalcolithic
Here, I will include two models. The first, is a minimal failure that does not include Steppe input.
left pops:
Arslantepe_LC
TellKurdu_EC
CHG
Iran_LN
numsnps used: 300549
best coefficients: 0.693 0.220 0.086
std. errors: 0.015 0.038 0.031
fixed pat wt dof chisq tail prob
000 0 13 25.383 0.0205459 0.693 0.220 0.086
This second model includes the Piedmont in the pleft, and is a success. Which may be more parsimonious beings that there is a legitimate steppe lineage there. While R1b-V1636 is certainly old enough to be part of the northern ancestry of CHG and there since the beginning, it is more in-line with data to withhold that assumption for the moment.
left pops:
Arslantepe_LC
TellKurdu_EC
CHG
Iran_LN
Piedmont_Eneolithic
numsnps used: 296251
best coefficients: 0.704 0.082 0.125 0.088
std. errors: 0.015 0.048 0.030 0.021
fixed pat wt dof chisq tail prob
0000 0 12 8.727 0.726049 0.704 0.082 0.125 0.088
Sources
Skourtanioti et al (2020) Genomic History of Neolithic to Bronze Age Anatolia, Northern Levant, and Southern Caucasus. https://www.sciencedirect.com/science/article/abs/pii/S0092867420305092
Population Genomics 