Global Admixture Analysis at K=6

December 31, 2013

This new study has one of the biggest global admixture analyses ever done (185 populations at 293,832 SNPs), but it's a little hard to view so I made a more user friendly version of one of the runs. I chose K=6 because it shows clearly the divisions and admixture between the six main racial groups: Capoid (red), Negroid (orange), Caucasoid (blue), Americoid (green), Mongoloid (yellow) and Australoid (purple). Click the image below to enlarge it and scroll down for examples of the racial types.


K=2 separates African from non-African populations.

K=3 reveals a West Eurasian ancestry component [Blue].

K=4 breaks the African component into an African hunter-gatherer ancestry maximized in Bushmen such as the Ju_hoan_North [Red] and an African farmer component maximized in the Yoruba [Orange].

K=5 breaks the ENA [eastern non-African] component down into one maximized in the Karitiana from the Americas [Green] and one maximized in the Ami from Taiwan [Yellow].

K=6 reveals a south Eurasian component maximized in Papuans [Purple], which is also represented in South Asians.


Ju'hoan (Namibia)
"Capoid"
Yoruba (Nigeria)
"Negroid"
Sardinian (Italy)
"Caucasoid"


Karitiana (Brazil)
"Americoid"
Ami (Taiwan)
"Mongoloid"
Papuan (PNG)
"Australoid"

12 comments

Truthteller said...

How is the purple component Australoid if the Onge aren't shown to be full purple?

Racial Reality said...

Obviously because it's almost 100% in Australian Aborigines and Papuans.

Truthteller said...

So then, are the Onge around 40% mongoloid? Because I have not read that anywhere. Yet that's what K=6 shows.

Anonymous said...

Nice one. I am a bit negative about that Iranians are so pure caucasoids tho'

Mark Royer said...

Are the Ainu missing from the study, or did I just skip over them?

roberto quintas said...

the study shows the different ethnics, not different races.

Varg said...

Sorry,but Papuan Austroloids probably are old Turkic Kypchaks.

Max said...
This comment has been removed by the author.
depescaun said...

roberto quintas, there are as many ethnicities as there are countries. This study obviously shows racial i.e. genetic differences. Seems to me like you've got it precisely backwards - you'd probably say that the French and the Germans are two distinct races, but same white ethnicity. They are two distinct ethnicities, but same white race.

Gregory76 said...

I long thought that we need to distinguish a Negritoid or Andamanoid primary race from the Australoid primary race, with Papuans (and Melanesians) being mainly Negritoid. They are distinguish phenotypically from Australoids by hair texture at least, being more like Negroids on this point.
Modern genetic work seems to me to confirm this, with modifications. It shows the two heavily mixed, with a large Australoid component in New Guinea and a large Papuan component in Australia. And it seems to show a third element.
This reminds me of Birdsell's trihybrid theory. He thought that there were 3 elements involved in Australasia: Tasmanoids (Negroid pygmies), Carpentarians (tall, dark and non-white Veddoids) and Murrayians (archaic Caucasoids Ainoids).
Y haplotypes and MtDna haplotypes seem to confirm this, mainly.
As to MtDna haployptes, Q is strong in New Guinea, N is strong in Australia and P is a strong minority in both. This suggests that Q is Negritoid, N is Australoid and P is Caucasoid.
As to Y haplotypes, M is strong in New Guinea, C is strong in Australia and K is a strong minority in both. This suggest that C is Australoid and K is Caucasoid, and would suggest that M is Negritoid, but M is descended from K, and so I think that the Caucasoids killed off the male husbands of Q and married the widows.
Also, when I say "Caucasoids" I mean people of Eurasian origin related to ancestors of Caucasoids, even thought there phenotype was probably had been shaped into something like that of the Negroids and Australoids.
And I would say that the Negritoids in the area need not have been of pygmy height. That seems to be an an adoption to dense forest.

Ane Koneo said...
This comment has been removed by the author.
Ane Koneo said...

This admixture analysis is slightly inaccurate and outdated for South Asians.

Other admixture analyses do not show South Asians' non-Caucasoid ancestry to be Australoid, and estimate them to have higher Caucasoid ancestry (by about 10-15%) than what’s shown here.

Yunusbayev et al. 2015 admixture shows that Pathans/Pashtuns are about 85% Caucasoid, Balochis are 90% Caucasoid, Brahmins and Kshatriyas (from Uttar Pradesh) are 75-80% Caucasoid and Kanjars (Rajasthani scheduled castes) are 75% Caucasoid. The admixture shows that their non-Caucasoid ancestry is East Eurasian (somewhat linked to Asian Mongoloids), not Australoid. https://i.imgur.com/Cvehgp7.png

Rasmussen et al. 2014 admixture shows the same, and also shows Sindhis are 80% Caucasoid, Burushos are 75% Caucasoid and Paniyas (Dravidian tribals) are 30-40% Caucasoid. It also shows that their non-Caucasoid ancestry is as linked to Australoids as the Mongoloids are linked to Australoids. https://i.imgur.com/sbIZURl.png

Karlsson et al. 2015 admixture shows that Gujaratis are 75-80% Caucasoid, and Bengalis are 55-60% Caucasoid. https://i.imgur.com/RSRl8a7.png

These estimates correlate better with phenotype. Yunusbayev and Rasmussen admixtures show that a South Asian component diverges at higher levels, but this component is still somewhat shared with Caucasus and Levant populations and not with Australoid populations. https://i.imgur.com/2eqCMiO.png Kusuma et al. 2016 admixture shows Indian Brahmins to be predominantly related to Europeans, and they do not diverge from Europeans even at K=14 (by which point several Mongoloid subgroups have diverged from each other). https://media.nature.com/lw926/nature-assets/srep/2016/160518/srep26066/images_hires/srep26066-f2.jpg