ADMIXTURE analysis of Eurasian populations with K=15

9 posts

I am a Noun

This is taken from Dienekes' Dodecad Ancestry Project

um, click on the link to get a better look.

SEU = Southern European
WAS = West Asian
CSI = Central South Asian
NEU= Northern European
NAF = North African
KRK = Korjak
WAF = West African
CHU = Chukchi
NGA = Nganasan
EAF = East African
ARB = Arabian
ALI = Altaic
CSA = Central Siberian


It begins.

I am a Noun

One thing I should point out and I didn't think about it until later is this is obviously group averages but individuals can be different if there is new admixture (like last thousand or so years). For example, I saw individual results for some Saudis who had no West African admixture, while some had excessive. Of course when you average it out you may get an impression of the group that isn't accurate for the individuals.

I also found it interesting that there was a Catalan with no North African influence but he's the only eastern Iberian I know for sure. Regardless, I've seen some stuff in the past that indicates that North African ancestry in Iberia goes from west (highest with Portuguese from this project having three times as much as the Spaniards on average) to east (lowest) instead of South (highest) and North (lowest).

Niccolo and Donkey
Lots of fun stuff like that: for instance Russians and Ukrainians with epicanthic folds that I have encountered.

Then we have all of those Central Asian groupings with some outliers in their appearances.

Let's see what Byssus comes up with when he catches this thread.
Team Zissou

True story: in the early 20th century, only "caucasian peoples" could get US citizenship. An immigrant from Syria to Atlanta, Costa Najour, sued for citizenship and won based on his Turkic/West Asian genealogy--an Assyrian. So Syrian immigrants became classified as "caucasian."

Bronze Age Pervert

I no longer believe in such studies. This one is considered a large study, but it contains only 1200 individuals TOTAL. Europe is a very mixed continent. If you pick individuals at random out of a population you will end up with a very skewed image of what's going on.

Cadavre Exquis
Not true. You will get a very precise picture of the population by using a large random sample.
Bronze Age Pervert
If the ruling class in a country is very small and ethnically distinct from the majority a small random sample might miss this distinction entirely. Similarly, if there are regional variations in racial groups as there are in France, you might miss this. Where are those subjects taken from, anyway? The cities? We don't know.
Cadavre Exquis
The subjects should ideally be picked from random, that was my point. As to variations within the population, you'd have to use distinct measures for selecting your sample (e.g. region, wealth). But even a small random sample can be reliable in describing the population as a whole .