I've done a supervised run on some of the samples I've been receiving. The set up was identical to a previous one, with three poles, Amerindian (unadmixed Totonac+Surui+Karitiana); Siberian (Unadmixed Nganasan); and Fertile Crescent (Egyptians). Results are basically identical, so I'm only publishing the new data:
I've named individuals with ancestry from one region of a country with a regional tag (N North, C Central, NW Nortwest etc). Individuals with mixed ancestry from a single country were given IDs with that country's name.
Some ancestry info:
CAsia is from Central Asia
Russian1 comes from North/Central Russia, Urals.
SEFI is a Swedish-speaking Finn
SwedenNorway has about 50/50 ancestry from either country
UKDK likewise UK and Denmark
CEuro has mostly German with some Polish and Ukrainian
SWFrance1 and CPortugal1 come from mountainous regions of their countries
PonticCaspian is mostly Gagauz with some North Ossetian
USA2 has major Amerindian ancestry
FRCA1 is French Canadian with colonial ancestry and also Faroe Islander
Brazil1 is a White Brazilian of colonial ancestry. No known recent Amerindian ancestors.
For some reason, just as in previous runs, if Fertile Crescent-like segments are pulled away by a suitable pole, Europeans with ancestry from the North or from mountainous regions have a small but detectable amount of DNA segments clustering with those of Amerindians. Other segments cluster with those of nearby Siberians. Samples from more Central/Southern or presumably fertile regions show such segments to a much smaller degree, being almost completely attracted to the Fertile Crescent pole.
I'm sorry for anyone not included this time, I didn't want to include too many recently mixed samples and also couldn't include individuals with major ancestry not covered by the poles. I'll run a more inclusive analysis soon.
Still accepting all non-recently admixed samples from the same country, even non-European ones. I have two Philippine samples already and would welcome more East, Southeast Asian ones.
I'm now considering more mixed samples as well as long as ancestry is well defined.