Continuing with the admixture analysis with our new reference 3 dataset.
Here's the results spreadsheet for K=13.
You can click on the legend to the right of the bar chart to sort by different ancestral components.
The Hadza were expected to split but I thought the San/Pygmy would split first.
Fst divergences between estimated populations for K=13 in the form of an MDS plot.
And the numbers:
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12
C2 0.093
C3 0.098 0.141
C4 0.179 0.212 0.192
C5 0.100 0.056 0.150 0.224
C6 0.112 0.149 0.075 0.210 0.153
C7 0.109 0.062 0.161 0.234 0.072 0.170
C8 0.181 0.222 0.208 0.279 0.232 0.226 0.239
C9 0.198 0.202 0.239 0.308 0.217 0.254 0.208 0.306
C10 0.164 0.186 0.145 0.276 0.184 0.146 0.217 0.290 0.303
C11 0.320 0.318 0.365 0.443 0.336 0.381 0.325 0.444 0.284 0.437
C12 0.261 0.263 0.302 0.377 0.277 0.318 0.270 0.371 0.153 0.370 0.278
C13 0.137 0.124 0.180 0.248 0.138 0.193 0.121 0.250 0.088 0.241 0.288 0.163
Is it possible for you to create an unrooted similarity tree of all the populations in your "Reference 3" dataset?
Yes I'll do that soon, though I should caution you not to think of it as unrooted tree since it is not a phylogeny.
Both Australian Aboriginals (Aus) and Great Andamanese should have their own component at higher Ks.
Is it possible for you to create an unrooted similarity tree of all the populations in your “Reference 3″ dataset?
An Fst spreadsheet of the Reference 3 populations would also be helpful.
In an Excel file.
Between population Fst can't be calculated with the Admixture program. It requires other software.
I didn't have in mind the ADMIXTURE software when I recommended Fst analysis for the populations.
There are many other softwares for genetic analyses.
I assume you want Fst for the actual populations since I do post the Fst for the ancestral components with every Admixture run.
Yes. I would be grateful if you do an Fst analysis of the actual reference 3 populations. Also, you can do an ASD (Allele Sharing Distance) analysis for them.
Can you point me to software that calculates ASD from plink files?
David (of Eurogenes) is much more knowledgeable about ASD than me, so I suggest you to ask him.
Also MCLUST and other MDS/PCA analyses.
Working on it.
Thanks.
The Hadza cluster is all over the place in Africa, makes little sense. The Mandenka and Yoruba should actually have close to 0% of it.
The Hadza are odd and seem to be very different than other African populations.
BTW, Zack, you can do the same analyses I recommended for the ref. 1 populations as well if you have time. Though this isn't as useful as doing them for the ref. 3 populations.