Genetic clusters of Turkic(and some Uralic) peoples, with new datas

[Sora]

Medieval Turkic percentage on modern Turkic populations;







1/2 Azeri from Kars, Turkey + 1/2 Oneida Native American



Salars(an Oghuz Turkic group live in Kansu, China)



Yugurs/Yellow Uyghurs(so called "pure" Uyghurs)



China Kyrgyz



Saami(an Uralic people in Northern Scandinavia)



Chuvash(seem to be predominantly Uralic)



Palestinian Turkmens (looks more Turkic than I've guessed)



Medieval Onoghur-Bulgar



Ancient Bulgar



China(East Turkestan/Xinjiang) Kazakhs



Orkhon Uyghur



Sheki Azerbaijanis



Zanjan(Iran) Afshars



Kazakhs from Central-West Kazakhstan



Turkmens from Shumanay(Uzbekistan)

[Sora]

Source: IG @turkic.ethnogenesis

[Komintasavalta]

Was that the best Chuvash beauty they could find...

If you try to model Chuvashes with all ancient population averages in the sources, they get a lot of Levänluhta:

$ mix <(aag|gn contam) <(mag chuvash) -p.02
Chuvash (.054):
74.9% FIN_Levanluhta_IA
11.0% RUS_Maykop
5.6% VK2020_NOR_North_VA_o1
5.2% VK2020_NOR_North_VA_o2
3.3% TUR_IA
$ mix <(aag|gn contam) <(mag mari,) -p.02
Mari (.080):
66.6% VK2020_NOR_North_VA_o1
25.7% FIN_Levanluhta_IA
6.4% RUS_Maykop
1.3% TUR_Tell_Kurdu_MC

But Chuvashes still get a much better distance if you add Maris to the sources like in the OP's image:

$ mix <(aag|gn contam;mag mari) <(mag chuvash) -p.02
Chuvash (.013):
66.6% Mari
16.6% VK2020_NOR_North_VA_o2
6.3% HUN_Prescythian_IA
3.0% Gepidian_SRB_ACD
2.9% RUS_Sunghir_MA
2.8% VK2020_UKR_Lutsk_MA
1.7% MDA_Cimmerian
0.1% TUR_IA

This uses my modified version of michal3141's convex optimization script: https://anthrogenica.com/showthread....l=1#post772891.

[Sora]

Bump