Admixture K=4,7,9, HRP0011 to HRP0020

We'll go to higher values of K (number of ancestral populations) for batch 1 later, but let's not keep the other batches waiting.

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 2 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 2 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 2 Admixture K=9

What do you guys think?

Higher values of K will be coming when admixture is done taking it sweet time to run. But more analysis and results are coming fast and furious now.


  1. Im 0015..I have 17% most curious about the southasian i will wait for higher Ks which can tell me more about the origins of my southasian part

  2. 14, 16, and 17, all South Indian Brahmins, seem to have similar admixture proportions.

    • an inbred lot u all 🙂 u guys look similar too ("tambram face"). what are you results when testing for run's of homozygosity?

      • Mine actually wasn't spectacularly high. (I'm HRP0017.) Without treating no calls as homozygous, I had 11 ROHs, with the longest being 1.10 Mb. 69.780% of autosomal SNPs were homozygous.

        Note that I'm on v3; the baseline will be lower for v2 people.

        (Would the other Tamil Brahmins -- 14, whom I don't know, in particular -- be willing to share any more ancestry info? Iyengar or Iyer, ancestral villages, that sort of thing?)

        • I should provide that info myself, since I'm asking. I'm Hebbar Iyengar (along with all my known ancestors). Note that we're Tamil-speaking, which might indicate migration from Tamil Nadu. (The other hypothesis is that Hebbar Iyengars were part of one of the mass Jain conversions in the 12th century, which seems less convincing to me.) Here's the ancestry info I can remember off the top of my head:

          * Father's mother's father's family was from Ambuga, though he was born in Shantigrama.
          * Father's mother's mother was from Belur.
          * Father's father's father's family was from Hampapura.
          * Father's father's mother's family was from Belur.
          * Mother's father's father's family was from Tippur.
          * Mother's mother was from Mysore.

          I'm also DOD331 ( on Dodecad and IN10 in Wesolowski's dataset. (And yes, we're basically identical on the ADMIXTURE runs and MDS plots there.)

          • Sri Vaishnavas migrating to Karnataka is no hypothesis, it's a historical fact through and through. Sri Ramanujacharya, with a big group of Sri Vaishnavas left for Karnataka due to the fact that the local Chola King, Kulothunga Chola II ordered for the forced conversion (and persecution of those who resisted) of Vaishnavas in to Shaivism, as he was a staunch Shaivaite himself. Upon Ramanuja's departure, it is said that the areas in and around his kingdom were plunged into darkness. The myth that some Hebbar Iyengars are in fact Jain converts developed in light of the fact that Bittideva the Hoysala King along with many other Jain nobles, converted to and became an adherent of Sri Vaishnavism upon the influence of Sri Ramanuja. He later came to be known as Vishnuvardhana. So essentially - Hebbar Iyengars are simply Iyengars who migrated to Karnataka - the major wave being with Ramanuja and minor other waves after this major one. One thing to be noted however, Hebbar Iyengars almost entirely belong to the Vadagalai distinction - which tells us that the Thengalais are a far more recent sub-sect. It seems almost obvious considering that the Thengalais emphasise the usage of the Tamil devotional literature over the Sanskrit Vedas! However, Iyengars themselves as an endogamous groups is quite a recent phenomenon.

          • It's not entirely clear during whose reign Ramanuja fled to modern-day Karnataka -- see the Wikipedia article, most of which I wrote a long time ago. :-p J.B. Carman argues convincingly that Ramanuja actually visited the melnadu twice. Regardless, Ramanuja's historical trips to Melkote don't prove that the ancestors of the Hebbar Iyengars came with him, any more than the Vānamāmalai JÄ«yar's journeys to North India in the early 20th century prove that the ÅšrÄ«vaiṣṇavas there migrated from Tamil Nadu.

            Neither is it obvious that the Jain origin hypothesis is a myth: there simply aren't sufficient contemporary records to answer the question definitively one way or another. Check out the discussion in Sastri's _History of South India_.

            The fact that Hebbar Iyengars are predominantly Vadagalai doesn't mean that the Tengalai school is a later innovation. The two tendencies developed roughly contemporaneously: for an overview, check out Patricia Mumme's definitive work on the subject, The Śrīvaiṣṇava theological dispute : Maṇavāḷamāmuni and Vedānta Deśika. There are people who argue that early Śrīvaiṣṇavism was more similar to contemporary Vadagalai practice, like Agnihotram Ramanuja Tatachariar, though most of the polemics on that front are pretty unconvincing.

          • At the risk of going way deeper into the weeds than anyone wants to go, I should also note that Hebbar Iyengars are hardly the only Iyengar group in Karnataka (even before the current period of varṇaÅ›aṃkara), and the other groups aren't all Vadagalai. B.L. Rice lists the following groups in 1897:

            Bhattacharyas (Tengalai), Embars (Tengalai), Hebbars (also known as Melnatar; either Tengalai or Vadagalai), Hemmigeyar (Vadagalai), Mandyattar (Tengalai), Maradurar (Tengalai), Metukunteyar (Vadagalai), Muncholi or Choli (Tengalai), Nallanchakravarti (Vadagalai), Prativaadibhayankarattar (Tengalai), Someshandal or Attan Kutattar(Vadagalai), Tirumaleyar (Vadagalai). Others not identified as either Vadagalai or Tengalai are Kadambiyar, Kandade and Kilnatar.

            And while Hebbar Iyengars are predominantly Vadagalai, there is a substantial Tengalai minority (some of whom I'm related to).

        • on v3, my fam, ROHs, autosomal, default settings, filter no-calls

          s1 - 69.735%, 14 runs, max 11.83
          s2 - 69.562%, 4 runs, max 1.78 mb
          rf - 69.785%, 12 runs, max 1.4 mb
          rm - 69.80%, 17 runs, max 1.92 mb

        • I'm HRP0014. Unfortunately I don't have much knowledge of my ancestral information, beyond being Iyer, and from the Tamil Nadu-Kerala border areas. I'm also relatively new to the entire ancestry assessment thing as well, so I've no idea how I'd do a run of homozygosity.

          Via 23andme, my maternal haplogroup's U2c, and my paternal haplogroup's H1a*, if that helps.

          • Figured out how to calculate it. Using Pike's tool, with settings on default:

            Treating no-calls as homozygous: 4 ROHs, 68.297% autosomal homozygosity.

            Without including no-calls, 4 ROHs, 68.012% autosomal homozygosity.

      • When setting "Treat No-Calls as homozygous when finding ROHs" to "No" I had 11 ROHs, with the longest being 1.82 MB. My total autosomal homozygosity is 69.714%. I am also on V3.

        I am HRP0016 for anyone who did not see my comments on prior posts.

        • Hi HRP0014,
          What is your Ancestry Painting at 23andMe like?

          Are you by any chance a Palagat Iyer? 💡

    • I am HRP13. Always suspected I had some Southeast Asian ancestry based on how some folks in my family looked. Cool!

  3. Gene Expression » Personal genomics around the web - pingback on February 10, 2011 at 2:04 pm
  4. You guys have no idea what inbred means. 😛

    If I don't count no-calls as homozygous, I get 147 ROHs on David Pike's tool with other settings on default.

  5. HRP13: 70.28% autosomal homozygosity

  6. Here are some versions of the spreadsheet sorted by the major component, which gives you an idea of where the breaks are in the data.

    • i have more kalash than southasian 13:17% but defintely to few to BE kalash

      • IIRC the European reference populations have the Kalash component. So you have inherited that component from both your European and your South Asian ancestry. That is probably why it is higher than your South Asian component, which only comes from your South Asian ancestry.

  7. Here is a bar-chart based on the spreadsheet sorted by geographic locale; west to east then south, its interactive.

    Go here

  8. Punjabi Brahmin HRP019

    Great barchart Mithra & Zack is doing just amazing work here! I am surprised to be missing any SE Asian and Papuan components. Did not see that coming. I wonder if the S Asian component would divide at some level (K=?) and what those subcomponents would mean - that is the biggest single source of curiosity for me. The Kalash component is also very intriguing. Zack, are you going to run HRP21-30 at K=9 also?

  9. Looking forward to a breaking into a ASI, ANI at higher levels of K!

    • Agreed! That would be very useful. Because as of now I see little sense in assigning a mixed type ("South Asian component") to a racially diverse people like South Asians.

    • I expect some split of South Asian but considering there is no ASI reference population according to Reich et al, I am not sure if we'll see ANI/ASI. But let's see.

    • Leaping recklessly to conclusions: Could it be that Kalash is ANI and S Asian is ASI? If so, an ASI split would be an original finding. Given how layered migrations have supposedly been (negrito, australoid, etc, etc), a split or several splits in s asian should happen.

      • Punjabi Brahmin HRP019

        Never mind. Just read that ANI ranges from 40-80%, so Kalash is not it - or not all of it anyway. S Asian should split to give both ASI and (part of) ANI. This will be so cool!

        • You answered your own question, I suppose, but yeah, "South Asian" should probably be understood as an ASI/ANI blend.

    • We should also be careful not to take the model posited by Reich, et al. as the last word. The apparent evidence of linear admixture they find could also be explained by simple isolation by distance, among other things.

  10. Fantastic bars and charts - thanks Zack and Mithra! I too would be interested in seeing a breakdown of the South Asian component. I know that my grandfather came from a village near Thelassery and that his mother was Malayali but I don't know my grandmother's South Asian origins.

    • If we don't see a breakdown of the South Asian component with the regular reference, we can exclude some out-groups and do a more South Asian focused analysis.

  11. So anyway, Zack - when are you going to post the next set of results? I'm particularly curious as to how HRP0029 would fare on your analysis - U.P Brahmins being the most populous Brahmins. He/She would essentially represent what could be deemed the "standard" Brahmin sample.

    • Working on batch 3 right now. Probably early next week.

    • I should be pretty close to the UP Brahmin, ie ~75-80% South Asian (S Asian + Kalash. Folk from the peninsula seem to have a higher value. For northern south asia that ~75-80% number seems to hold true.

      • If the U.P Brahmin individual is willing to reveal his haplogroups here - please go ahead. I am very curious as to what they are.

          • Thanks for sharing! 😀
            I'd have personally expected a U.P Brahmin individual to be R1a1 considering it's high frequency among them, but H1 doesn't come as a surprise either as U.P Brahmins carry that at a frequency of 16.13%.

            I wish this project had more U.P "Lower Caste" samples. Considering it's the most populous state in India, an admixture analysis on U.P lower caste individuals would prove to be useful for reference..

  12. New here.. just got the email from 23andme.. paternal haplogroup is R1a1a and maternal haplogroup is U2a

    no expertise in genetics.. but looking to learn

    tamil iyer

  13. @ RK

    It is widely accepted in TN & A.P that Vadakalais are pure blood(indo-aryan) while thenkalais have admixed with non-brahmin sudras(tn non brahmins are dravidians). It is too obvious from their appearance. Thenkalais do have dravidian admixture.
    Much of the thenkalais in triplicane are believed to be of mandyam origin. And many, including iyers and madhwas maintain that these thenkalais have admixed with a large number of sudras.

    Also, a vadakalai matrimony profile will always read "alliance needed from our sect/vadakalai only", while that in a thenkalai profile is always "Kalai No Bar".

    - I'm talking about the mainstream iyengars in Tamil Nadu & some of those in A.P.

    I don't know much about hebbars, however ethnographers have always maintained that they are "brahmin migrants from tamil nadu". Anthropologists never spoke of any jain factor.

    Also, apart from that list "hebbar, mandayattar etc" that you've suggested, a majority of iyengars (mainstream from tamil nadu)are just called vadakalais & thenkalais, who never fall into any of those sub-categories. Those categories that you suggested are just minorities. There are just pancharatras & munitrayas among the mainstream vadakalais in TN.

Trackbacks and Pingbacks: