running these samples (less the three outliers) together with the K13 updated sheet:
Target: Nurzat
Distance: 0.3427% / 0.34270312
39.3 Ukraine_Ternopilska
17.7 Ukraine_Kropyvnytska
12.8 Ukraine_Sumska
7.2 Abhkasian
7.2 Ukraine_Zakarpatska
6.7 French_Basque
3.3 Sardinia
2.8 Dargin
1.3 Greek_Northern-Thrace
0.7 Turk_East_Black_Sea
0.4 Yoruban
0.3 Dai
0.2 Greek_Western-Thrace
0.1 She
then, only using the Ukrainian samples and Romanian 10 regions averages:
Target: Nurzat
Distance: 1.7171% / 1.71713205 -
distance too high
53.3 Ukraine_Ternopilska
15.3 Romania_Muntenia
13.2 Romania_Oltenia
10.2 Ukraine_Zakarpatska
5.5 Ukraine_Kropyvnytska
2.5 Ukraine_Chernivetska
when to Ukrainian and Romanian I add Greek and German samples - much better distance (Caucasus-rich Greek makes sense in the case of Greek boyars in Moldova, that often actually came from present Turkey), but funnily no Romanian samples get chosen, maybe because Moldovans are naturally a sort of a Ukrainian-Greek mix (not ethnically or recent but in terms of macro components, like Farmer plus Steppe model)?
Target: Nurzat
Distance: 0.9940% / 0.99401827
52.4 Ukraine_Ternopilska
21.6 Ukraine_Zakarpatska
11.3 Greek_Western-Thrace
5.3 Greek_Trabzon
5.1 German_West
4.3 Ukraine_Kropyvnytska
Distance to: Nurzat