Admixture K=4,7,9, HRP0021 to HRP0030

Here's the spreadsheet with their admixture results. And you can check their ethnic backgrounds.

You might also want to refer to the reference dataset I admixture analyses for K=2-5 and K=6-9.

I did not run admixture for all values of K this time. So let's start with K=4. For quick reference,

C1 South Asian
C2 European
C3 East Asian
C4 African

Batch 3 Admixture K=4

Now, for K=7, the ancestral components are:

C1 South Asian
C2 European
C3 Southeast Asian
C4 Southwest Asian
C5 Papuan
C6 Northeast Asian
C7 African

Batch 3 Admixture K=7

And finally, here's K=9.

C1 South Asian
C2 Kalash
C3 Southwest Asian
C4 Southeast Asian
C5 European
C6 Papuan
C7 Northeast Asian
C8 West African
C9 East African

Batch 3 Admixture K=9


  1. As expected the UP Brahmin (HRP0029) and Bihar Brahmin (HRP0003) are almost exactly the same.

  2. HRP0025, "Karnataka Brahmin" is pretty similar to the other South Indian Brahmins from the previous batch. Additionally, HRP0024 "Andhra (Hyderabadi)" and HRP0026 "Goan Catholic Brahmin" look similar to us as well.

  3. At K=9, in terms of the admixture percentages, the following are the closest pairs:

    HRP0024 & HRP0025
    HRP0014 & HRP0017
    HRP0002 & HRP0023 (child & parent)
    HRP0016 & HRP0026
    HRP0017 & HRP0024

  4. A graph with all the participants, sorted by geographical locale West to east then south

    Go here

  5. Thanks for providing the interactive map, Mithra.

  6. HRP019 Punjabi Brahmin

    Now I am beginning to wonder what the NE Asian signifies. Where would it come from exactly?

  7. Now I am beginning to wonder what the NE Asian signifies. Where would it come from exactly?


    notice that bengalis have mostly southeast asian, with a minor east asian. northwest indians who have east asian are almost all northeast asian. while southern indians' east asian tends to be southeast asian.

    • It could also be from Mongols and Turkic populations.

      • yes, but by the time these groups came to s asia in large numbers they were muzzie. so we're positing hinduisation. not impossible. but it seems higher probability that you might have a hindu from nepal with tibetan ancestry being integrated into north indian brahmin jatis.

        • HRP019 Punjabi Brahmin

          it would be good to look at some tibetan dna. if tibetans are a mix of se asian and ne asian (like the chinese seem to be), then it is likely not tibetan, don't you think? else there would also be an se asian component there. maybe it is huna in origin - shakas or whoever else - that must have had some mongol/turkic/altaic element in it. all this is raising many more questions in terms of its implications than it is answering.

          -why do the jatts have a notch more european in them than the iranians?

          -what is the papuan representing? which group in india has it the most? the andamanese? which group on mainland india? the bhils? gonds?

          -is a split actually likely in south asian? i mean sw asian has already split from kalash and w african from e african. s asian feels pretty stubborn. i'd be curious to see what splits upto, say, k=12. maybe sw asian will split first 🙂

          -where is the south asian 7-8% in iranians coming from?

          this is all so tantalizing. can mithra create another interactive using selections from your other reference set combined with the sets here? something that includes gujaratis, rajputs, sindhis, balochis, tibetans, pathans, marathis?

          • I should also admit to something else here. Like many Punjabi and Kashmiri types, we too have had people of other faiths absorbed into our family. In my case, one of my great-grandmothers (my maternal grandfather's mother). However, she was of similar ethnic background as his father - so I did not flag this to Zack when I sent him my data.

  8. Looking at Mithra's figures I noticed two things about the Indo-Caribbeans - one is that they're basically Tamils, and the other was that one of them had 4% West African ancestry.

    While I'm not the least surprised to see West African ancestry in Indo-Caribbean people, I was a little bit surprised at how small the component was. Not having poked around all that much, I don't have a sense of how 'West African' someone who actually was West African would appear in an analysis like this, but assuming it's in the 50-100%, this would represent a great-great or great-great-great grandparent. While generation times vary among families, I don't think a lot of us have Indian great-great-great grandparents already in the Caribbean. It's far from impossible (my cousin's 30-year-old kids would have that many generations here), but certainly put at least one of their ancestors among the earlier immigrants.

    • That might be a possibility. Or one of the ancestors might have been mixed (European & African? Native American & African? Etc)

      • It's possible, but it only pushes it back a generation, I'd think. The person shows up as 5% European, which fits in with the other south Indian samples. Of course the person right above them in Mithra's figure (Tamil, Nadar caste) has none. So it's also possible that this person's European blood* comes from the same place as their African ancestry.

        As for Amerindian - would that not show up with this PC?

        *i'm struggling with language here. I realise that the percentages here don't necessarily reflect ancestry, but rather SNP patterns of similarity.

        • Bear in mind that the samples from both North India and South India are pretty unrepresentative of the actual demographics there: Nadars make up about 12% of the population of Tamil Nadu, and the Brahmin groups make up around 1-4%. And obviously huge stretches of North India (and several communities there) aren't represented either. You can see here that they're just about as similar to some of the Houston Gujaratis as they are to the South Indian participants. So I'm not sure how certain we can be about where their Indian ancestors came from (and thus, how consistent their results are with post-arrival West African admixture).

          Maybe a PCA plot or MCLUST would reveal more.

          • The West African admixture seems certain and it is much more likely to be post-arrival in the Caribbean. The other details are of course speculation on our part.

          • Yeah, I was referring to whether the West African ancestor's admixture with a non-West African (either European, Native American, or East Indian) took place before or after the arrival of the Indian ancestors. PCA or MCLUST might help reveal where the Indian ancestors came from.

          • The thought crossed my mind that the African ancestry might have come from India (given that there are populations in India with African origins) but it seemed more likely that these people would have had East African, rather than West African ancestry (though, of course, that wouldn't hold if Bantu-speakers show up as West Africans, rather than East Africans).

            That said, it still seems more likely that African ancestry was picked up in the Caribbean. Knowing where in the Caribbean the person came from would and how long their family had been there would make one model or the other more (or less) likely.

          • East African Bantus do show up as majority West African at K=9 admixture, though they separate at K=12.

        • Since I am not using an Amerindian samples, Amerindian ancestry will show up as East Asian.

    • Aren't Indians in the English-speaking Caribbean mostly the descendants of people from the Indo-Gangetic plain (i.e., Uttar Pradesh and Bihar)?

      • Most of the immigrants departed from northern ports, but there was a substantial minority who departed from Madras. There has always been a substantial south Indian element in Trinidad. I don't know the actual proportions (but it would be easy enough to check the ship records, see where they departed from) but I feel it's something like 1/3.

  9. Yes, I was surprised by this too, though caste may account for differences between UP/Bihar Brahmins and them.

    Also, UP and Bihar are heterogeneous populations. West UP is more like Haryana with similar communities like Jats, etc. West Bihar is more like East UP. Jharkhand has a strong tribal element and so on.

    Zack's hypothesis on the source of WAfr percentage makes sense to me.

  10. Admixture K=12, HRP0021-HRP0030 | Harappa Ancestry Project - pingback on February 24, 2011 at 1:26 am

Trackbacks and Pingbacks: