Continuing with the admixture analysis with our new reference 3 dataset.
Here's the results spreadsheet for K=8.
You can click on the legend to the right of the bar chart to sort by different ancestral components.
Fst divergences between estimated populations for K=8 in the form of an MDS plot.
And the numbers:
C1 C2 C3 C4 C5 C6 C7
C2 0.098
C3 0.073 0.139
C4 0.090 0.152 0.065
C5 0.184 0.201 0.220 0.231
C6 0.113 0.068 0.147 0.166 0.223
C7 0.164 0.208 0.180 0.170 0.273 0.227
C8 0.158 0.139 0.181 0.212 0.285 0.143 0.275
Who are Aus?
Australian aboriginals.
They have too much from the Mongoloid and South Asian components. I wonder whether those components are almost totally post-colonial like the European and SW Asian components.
The East Asian cluster found in the Australians probably isn’t really East Asian but something Australoid which is different from the Papuan component. Perhaps when the Bouganville Melanesians create their own cluster it will decrease significantly.
Ez, yes, and that probably also includes the South Asian and Siberian components, as I wrote in my comments below.
By "post-colonial" I mean post-European colonialism.
almost totally post-colonial like the European and SW Asian components
Actually the West Eurasian components in Australian Aboriginals should be totally post-colonial, not almost totally. But when it comes to the East Eurasian components a tiny segement of them may be expected as in Papuans.
segment
You can test whether the components other than the "Papuan" component in Aus are totally or almost totally post-colonial by looking at their distribution in individual Aus samples (looking at how homogeneously or heterogeneously distributed they are).
There are three samples. One of them has significant Admixture: 7, 20, 23, 3, 47, 0, 0, 0. Probably only half aboriginal.
The average for the other two is 10, 8, 0, 0, 81, 1, 0, 0.
The average for the other two is 10, 8, 0, 0, 81, 1, 0, 0.
Assuming they are unadmixed Aboriginals, either pre-colonial Aboriginals were more East Asian and South Asian admixed than Papuans, or they had a gene pool that is very divergent from Papuans and didn't form their own component (at least not at such low Ks)due to the low number of Aboriginal samples (just three).
they had a gene pool that is very divergent from Papuans and didn’t form their own component (at least not at such low Ks) due to the low number of Aboriginal samples (just three)
I mean they may have been so divergent from Papuans that they may be being represented in ADMIXTURE analysis at these Ks as a mixture of a major Papuan component with a low proportion of the East Eurasian and South Asian components despite not having actual East Eurasian and South Asian admixture as a result of low sample size.
BTW, Zack, are you planning to publish (in this blog or in a medium like Rapidshare) the ADMIXTURE results of the reference populations on an individual by individual basis like Dienekes?
I haven't seen Dienekes do that. But sure I can upload those if anyone's interested. Are you interested in a specific K or all K? I can zip it up and upload along with an info file. Can't prettify them like I do with the population average results or project participant results.
Yes, Dienekes did that under the label "Portraits". I prefer all Ks as every K tells a different part of the whole picture.
A spreadsheet would be just fine. Preferably all K. 🙂
I'll upload them over the weekend.
Thanks. That will certainly help make better sense of the reference population results.
Good idea Onur. I would be interesting in seeing the individual results as well.