0
Forgot to mention that this is what I was told why can't rely on Admixture, PCA or G-25 for actual direct ancestry from a specific population is because they cluster based total drift since the beginning of time. For example, if we look at 3 subjects on PCA or G-25
an Italian
a British
a half British half Ethiopian mulato
The British and Italian will cluster near each other or on G-25 have closer coordinates to each other than the Mulato to the British, even though the Mulato has 1 actual British parent because the British and Italian share a ton of very very old ancestry whereas the Mulato has 50% African ancestry which would place him further away from the European cluster even though 1 of his parents is actually British and the Italian doesn't have a British parent.
That's why PCA or G-25 is not accurate when it comes to more recent direct ancestry especially if we try to find steppe ancestry in Middle-Eastern populations that have alot of Basal Eurasian ancestry. The Basal Eurasian ancestry plots them further away from Andronovo than a European on PCA or G-25 since Europeans and Andronovo share more much older ancestry with each other (more WHG and EHG) than a Middle-Eastern and Andronovo (less WHG less EHG more Basal more SW Asian )
That's we have to do IBD using a reliable method or use some formal stats to figure out more recent direct ancestry. That's also why PCA and G-25 will change alot depending on what samples are in the run. Doesn't mean that IBD is 100% foolproof since it has little error margin but it's generally much more suitable for figuring out direct ancestry
Bookmarks