What the Harappa Ancestry Project has resolved

By Razib Khan | August 4, 2013 7:39 pm

My friend Zack Ajmal has been running the Harappa Ancestry Project for several years now. This is a non-institutional complement to the genomic research which occurs in the academy. His motivation was in large part to fill in the gaps of population coverage within South Asia which one sees in the academic literature. Much of this is due to politics, as the government of India has traditionally been reluctant to allow sample collection (ergo, the HGDP data uses Pakistanis as their South Asian reference, while the HapMap collected DNA from Indian Americans in Houston). Of course this sort of project is not without its own blind spots. Zack must rely on public data sets to get a better picture of groups like tribal populations and Dalits, because they are so underrepresented in the Diaspora from which he draws many of the project participants.

Once Zack has the genotype one of the primary things he does is add it to his broader data set (which includes many public samples) and analyze it with the Admixture model-based clustering package. What Admixture does is take a specific number of populations (e.g. K = 12) and generate quantity assignments to individuals. So, for example individual A might be assigned 40% population 1 and 60% population 2 for K = 2. Individual B might be 45% population 1 and 55% population 2. These are not necessarily ‘real’ populations. Rather, the populations and their proportions are there to allow you to discern patterns of relationships across individuals.

Since Zack has put his results online, I thought it would be useful to review what patterns have emerged over the past two years, as his sample sizes for some regions are now moderately significant. Though he has K=16 populations, not all of them will concern us, because South Asians do not tend to exhibit many of the components. I will focus on seven: S Indian, Baloch, Caucasian, NE Euro, SE Asian, Siberian and NE Asian. These are not real populations, but the labels tell you which region these components are modal. So, for example, the “S Indian” component peaks in southern India. The “Baloch” in among the Baloch people of southeastern Iran and southwest Pakistan. The “NE Euro” among the eastern Baltic peoples. The last three are Asian components, running the latitude from south to north to center. They only concern the first population of interest, Bengalis.  I will combine these last three together as “Asian.”

Below is a table, mostly individuals from Zack’s results (though there are some aggregate results from public data sets). Comments below.

Bengali Baidya45%30%3%5%12%
Bengali Baidya45%27%3%6%12%
Bengali Brahmin45%35%2%11%4%
Bengali Brahmin44%35%5%11%4%
Bengali Brahmin43%35%4%10%4%
Bengali Brahmin42%32%4%8%6%
Bengali Brahmin41%33%7%8%5%
Bengali Brahmin40%33%4%10%4%
Bengali Brahmin40%30%6%10%7%
Bengali Muslim50%25%1%5%15%
Bengali Muslim49%28%3%4%15%
Bengali Muslim45%27%4%4%17%
Bengali Muslim45%26%2%2%16%
Bengali Muslim45%24%1%3%19%
Bengali Muslim43%25%3%2%18%
Bengali Muslim48%27%0%5%15%
Tamil Brahmin48%37%6%5%
Tamil Brahmin48%37%3%5%
Tamil Brahmin48%35%5%6%
Tamil Brahmin47%38%6%4%
Tamil Brahmin47%40%3%5%
Tamil Brahmin46%40%3%6%
Tamil Brahmin Iyengar50%35%2%8%
Tamil Brahmin Iyengar47%38%6%4%
Tamil Brahmin Iyengar47%35%6%6%
Tamil Brahmin Iyer48%38%4%5%
Tamil Brahmin Iyer48%38%2%5%
Tamil Brahmin Iyer47%37%2%5%
Tamil Brahmin Iyer47%37%6%8%
Tamil Brahmin Iyer43%35%6%5%
Tamil Muslim58%28%3%2%
Tamil Nadar62%30%0%0%
Tamil Nadar59%32%3%0%
Tamil Nadar55%30%3%0%
Tamil Vellalar50%35%6%1%
Tamil Vellalar51%32%5%0%
Tamil Vellalar (Sri Lankan)60%32%5%0%
Tamil Vellalar (Sri Lankan)60%33%0%0%
Tamil Vellalar (Sri Lankan)56%36%0%0%
Tamil Vishwakarma70%23%0%0%
Tamil Vishwakarma66%25%4%0%
Andhra Pradesh60%34%2%0%
Andhra Pradesh54%36%2%3%
Andhra Pradesh (Hyderabad)56%29%5%0%
Andhra Pradesh (Hyderabad)47%35%8%4%
Andhra Pradesh Gouda61%30%2%1%
Andhra Pradesh Kamma51%33%7%0%
Andhra Pradesh Kapu62%30%2%1%
Andhra Pradesh Naidu51%32%4%2%
Andhra Pradesh Reddy57%37%1%0%
Andhra Pradesh Reddy54%38%3%0%
Andhra Pradesh Reddy51%35%4%0%
Andhra Pradesh Reddy50%36%2%1%
Andhra Pradesh Telegu Brahmin45%33%6%4%
AP Brahmin (Xing, N = 25)49%36%3%6%
AP Naidu (Reich, N = 4)61%31%1%1%
Kannada Devanga60%31%3%1%
Karnataka Catholic Christian56%37%3%0%
Karnataka Lingayat55%34%4%0%
Karnataka Brahmin51%35%3%5%
Karnataka Iyengar49%36%5%5%
Karnataka Iyengar48%39%3%5%
Karnataka Iyengar48%37%3%7%
Karnataka Brahmin47%38%4%6%
Karnataka Konkani Brahmin47%37%2%6%
Karnataka Konkani Brahmin46%33%6%7%
Karnataka Kokani Brahmin44%34%6%5%
Kerala Brahmin43%39%4%6%
Kerala Christian53%35%4%0%
Kerala Christian50%35%8%1%
Kerala Christian45%33%7%3%
Kerala Muslim Rawther53%35%2%1%
Kerala Muslim Rawther51%28%4%3%
Kerala Nair48%40%4%0%
Kerala Nair47%38%5%5%
Kerala Syrian Christian50%37%6%0%
Kerala Syrian Christian50%35%9%1%
Kerala Syrian Christian46%33%5%4%
Kerala Syrian Christian44%33%6%4%
Pathan (HGDP, N = 23)23%42%16%11%
Kalash (HGDP, N = 23)22%43%18%11%
Burusho (HGDP, N = 25)23%41%12%10%
Brahui (HGDP, N = 25)12%58%12%2%
Sindhi (HGDP, N = 24)29%46%10%6%
Kashmiri Pandit (Reich, N = 5)32%39%12%9%
Punjabi Arain (Xing, N = 25)31%44%10%7%
Punjabi Brahmin35%40%8%11%
Punjabi Brahmin33%41%13%10%
Punjabi Chamar40%33%9%6%
Punjabi Jatt28%39%11%10%
Punjabi Jatt30%44%6%14%
Punjabi Jatt28%42%8%13%
Punjabi Jatt28%46%7%13%
Punjabi Jatt28%40%10%15%
Punjabi Jatt27%44%10%13%
Punjabi Jatt27%35%16%11%
Punjabi Jatt Muslim30%39%13%8%
Punjabi Khatri30%42%12%12%
Punjabi Lahori Muslim31%44%11%8%
Punjabi Pahari Rajput34%43%11%7%
Punjabi Pakistan28%36%16%7%
Punjabi Ramgarhia35%43%5%9%
Haryana Jat25%33%12%17%
Haryana Jat25%33%12%17%
Haryana Jatt28%38%5%20%
Haryana Jatt26%39%10%17%
Rajasthan Marwari Jain47%34%5%6%
Rajasthani Agarwal51%37%6%1%
Rajasthani Brahmin32%38%9%15%
Rajasthani Marwari48%34%6%2%
Rajasthani Rajput45%38%5%9%
UP Brahmin41%37%7%11%
UP Brahmin40%37%7%11%
UP Brahmin37%38%2%14%
UP Kayastha47%38%5%3%
UP Muslim33%33%10%9%
UP Muslim28%35%12%11%
UP Muslim Pathan48%36%7%4%
UP Muslim Syed33%31%13%7%
UP Syed36%37%7%8%
UP/Haryana Agarwal52%35%6%2%
UP/Haryana Jatt28%42%7%18%
UP/Madhya Pradesh51%27%1%7%
UP/Punjabi Khatri27%43%10%11%
Bihari Baniya47%31%5%5%
Bihari Brahmin39%38%5%11%
Bihari Kayastha53%33%1%7%
Bihari Muslim48%28%5%8%
Bihari Muslim42%34%9%6%
Bihari Muslim41%36%7%8%
Bihari Muslim42%32%7%9%
Bihari Syed42%35%4%9%
Gujarati (HapMap, N = 63, Patel)54%42%0%1%
Gujarati (HapMap, N = 34, Non-Patel)44%39%5%7%

A recent paper suggested that there was a single pulse of admixture between South and East Asians in the environs of what is today Bangladesh which occurred ~500 A.D. The traditional accounts for the arrival of Brahmins to Bengal suggests a period around and after 1000 A.D. (Bengal was one of the last redoubts of institutional Buddhism in northern India, so presumably would have less need for the services of Brahmins). The results are easy to align with these two facts. All the Bengali non-Brahmins (Baidya are a non-Brahmin high caste in West Bengal) have substantial East Asian ancestry. The Bengali Brahmins have far less of this. Additionally, their “NE Euro” component is about double that of non-Brahmins. There is still room for the Bengali Brahmins being a synthetic community with some admixture (their East Asian fraction is still notably higher than elsewhere in South Asia), but the outlines of the traditional narrative seem to explain the broad outline of these results.

When you look at South Indians from the four Dravidian states there are four facts which strike me as of note:

– There is a distinct difference between Brahmins and non-Brahmins (most of the non-Brahmins Zack has in the Harappa data set are upper caste, though the public data sets have Dalits and tribal populations)

– There is very little difference between South Indian Brahmins by region and sect (e.g., Iyengar vs. Iyer are Tamil Brahmins divided by theological differences).

– South Indian Brahmins are genetically distinct from North Indian Brahmins. They seem to have about one half the proportion of the “NE Euro” component as North Indian Brahmins (e.g., compare to Bengali Brahmins).

– South Indian non-Brahmin upper castes have very little of the “NE Euro” component, which is found at low, but consistent fractions among non-Brahmins in the Gangetic plain (and at much higher fractions as one moves toward the Punjab)

I do not know about the nature of the origin of the Pancha-Dravida group of Brahmins, but they look to be endogamous, from the same source, and probably had some admixture with the local substrate early on. This would explain their uniformity and lower fraction of “NE Euro” in relation to North Indian Brahmins. The results above also suggest that the Syrian Christians derive from converts from the Nair community, or related communities. This should not be surprising.

Finally let’s move to North India, and the zone stretching between Punjab in the Northwest and Bihar in the East. Though in much of this region Brahmins have higher “NE Euro” fractions, this relationship seems to breakdown as you go northwest. The Jatt community in particular seems to have the highest in the subcontinent. There are inchoate theories for the origins of the Jatts in Central Asia. I had dismissed them, but am thinking now they need a second look. The reasoning is simple. The Jatts of the eastern Punjab have a higher fraction of “NE Euro” than populations to their northwest (Pathans, Kalash, etc.), and Brahmin groups (e.g., Pandits) in their area who are theoretically higher in caste status. This violation of these two trends implies something not easily explained by straightforward social and geographic processes. The connection between ancestry and caste status also seems to break down somewhat in the Northwest, as there is a wide variation in ancestral components.

Someone with more knowledge of South Asian ethnography should weigh in. But until then I invite readers of South Asian heritage to submit their results to Zack.


Comments (14)

  1. gpandatshang

    Do we have any hypotheses about the historical events behind the Bengal hybridization? Is it reasonable to surmise that it was related to the Völkerwanderung that brought the Thaioid and Burmesque peoples into Southeast Asia?

    I wonder if the SE Asian hybridization could have influenced Bengal’s Buddhist inclinations. I had speculated in a comment a few months ago that perhaps Buddhism was never really a mass religion in South Asia. One thing we can say for certain is that today, Southeast Asian countries today are thoroughly Buddhist. Perhaps popular Buddhism was introduced to Bengal from Southeast Asia at a relatively late date (ca. 500 CE would seem relatively late by Indian standards but surprisingly early by SE Asian standards).

    • razibkhan

      too late. buddhism shows about 500 years later in SE asia in force. it is of the theravada variety, not the tantric style more in vogue in bengal (which influenced tibet). also, the idea of a ‘mass religion’ may simply not be appropriate for that period in any case. until recently most ‘hindus’ were simply not muslim/christian/jew/zoroastrian. they did not practice or know of the religion of adi sankara, for example (general insight, not just hinduism);

      • gpandatshang

        Not necessarily. Southeast Asia today is uniformly Theravadin (given that, in terms of Buddhist culture, Viets count as East Asian). However, there seems to have once been a substantial Mahayana influence, and I believe one can still find occasional veneration of Avalokiteśvara there. Regarding the former, per Karuna Kusalasaya (http://www.accesstoinsight.org/lib/authors/kusalasaya/wheel085.html — I don’t know the author specifically, but the site is usually reliable and well-edited):

        “With the growth of Mahayana Buddhism in India, especially …. during the second half of the first century A.D., the sect also spread to the neighboring countries, such as Sumatra, Java, and Kambuja (Cambodia). It is probable that Mahayana Buddhism was introduced to Burma, Pegu (Lower Burma) and Dvaravati (now Nakon Pathom in Western Thailand) from Magadha (in Bihar, India) at the same time as it went to the Malay Archipelago. But probably it did not have any stronghold there at that time; hence no spectacular trace was left of it. Starting from the beginning of the fifth century A.D. Mahayana Buddhist missionaries from Kashmir in Northern India began to go to Sumatra in succession. From Sumatra the faith spread to Java and Cambodia.”

        As for tantra, I doubt there was ever a clear distinction made between tantric Mahayana and sutric Mahayana outside of East Asia, but in any event Wikipedia has a whole article on “Vajrayana Buddhism in Southeast Asia”.

        I think it would be reductivist to say that “mass religion” is simply an inappropriate category for this time/place. Even though the peasant public has never been “theologically correct”, they still had ties (albeit sometimes complicated or conflicting) to various religious establishments. For instance, a medieval Burmese peasant who felt a need for good luck or for exorcism would likely look for a bhikkhu, whereas a peasant from a “Hindu” part of India probably would not.

        I know the above is speculative; my point is just that I can’t rule out the possibility that Bengal was proselytized by Mahayana Buddhists from Southeast Asia ca. 500 CE, perhaps accounting for a more robust popular form of Buddhism there.

        I’m more interested in your thoughts on the historical events that might lie behind the single pulse of admixture with South and East Asians.

        • razibkhan

          I think it would be reductivist to say that “mass religion” is simply an inappropriate category for this time/place.

          what does reductivist mean here?

          fwiw, the other area of strength of buddhism in northern india was punjab on the eve of islamic eruption.

          I’m more interested in your thoughts on the historical events that might lie behind the single pulse of admixture with South and East Asians.

          now that i think a folk wandering from the east seems plausible. though probably tibeto-burman, not thai.

          • gpandatshang

            By “reductivist”, I just meant an oversimplification, Mass religion on the Protestant model (i.e. in terms of orthodox belief) is a problematic idea, but it would be an exaggeration to say there is no such thing; it’s just that it’s fuzzy and complicated (perhaps more so wrt indigenous Indian religions than for Christianity & Islam).

            As for the folk wandering, yes Tibeto-Burman seems much more likely; Daic would have been later. WIkipedia claims there were Tibeto-Burman speakers in Burma in the 2nd century BCE, but that could easily be nationalistic overstatement (something that Wikipedia very vulnerable to).

          • razibkhan

            of course i didn’t mean the protestant model, that’s atypical even within xtianity. even conventional islamic orthopraxy as a mass religion is a new thing (see: rural bengali muslim reforms early 20th century). not sure how i can actually get a sense of what you mean though…because i don’t know the outlines of your argument aside from the one sentence model you presented. confessional adherence in a very deep-in-the-bone sense in europe didn’t seem to kick in until the 17th century (this is when changes in the religion of rulers would start to result in resistance from the broader population; e.g. conversion to calvinism by the hohenzollerns).

            WIkipedia claims there were Tibeto-Burman speakers in Burma in the 2nd century BCE, but that could easily be nationalistic overstatement (something that Wikipedia very vulnerable to).

            the easiest way to check is have zack check the bengalis in a small supervised run in ADMIXTURE. i will ask.

          • “confessional adherence in a very deep-in-the-bone sense in europe didn’t seem to kick in until the 17th century (this is when changes in the religion of rulers would start to result in resistance from the broader population; e.g. conversion to calvinism by the hohenzollerns).”
            That got me wondering when the “Old Believer” split occurred, and sure enough its the 17th century. There have been older divisions regarding theology such as the Catholic vs Monophsite dispute which gave rise to the “Greens” and “Blues” of the Nika riots, but I don’t know if that’s comparable.

          • gpandatshang

            My point was to wonder how Bengal got to be one of the last redoubts of institutional Buddhism in northern India. One hypothesis would be purely élite-oriented: Bengal had more late Buddhist kings, who empowered a Buddhist religious establishment. I was wondering if an alternate hypothesis could be that the Bengali public had a different attitude toward the Buddhist establishment than other north Indians did, with more popular involvement and attachment. You said that the concept of mass religion may simply not be appropriate for that period, but I disagree: complicated, yes, but I don’t see any reason to think “mass religion” is inapplicable. I didn’t mean to imply that you meant the Protestant model; I was cautioning against going too far in the opposite direction, to suppose that religious establishments mean nothing to the public – I guess this is the salient distinction: popular dedication to doctrine might be rare/late, but popular ties to particular religious establishments (such as the order of Buddhist monks) could be robust without dedication to doctrine.

            To draw a connection between Bengali Buddhist identity and the hybridization event of ca. 500 CE is pure speculation but entirely plausible.

          • razibkhan

            complicated, yes, but I don’t see any reason to think “mass religion” is inapplicable.

            the domain of popular religion is relatively thin. but, there is some information on china, and to a lesser extent the west (christian + islam inclusive). i see plenty of reason to be highly skeptical of the model you’re presenting, if by ‘popular’ you mean peasant cultivators (as opposed to local sub-elites), as someone who knows more about the history of religion than almost anyone i have met (i’m not bragging, i’m being truthful). it’s your freedom to disagree. but just asserting your disagreement doesn’t really forward the discussion. it’s hard to motivate someone to falsify a contention when it relies on an axiom that they reject on prima facie grounds.

          • gpandatshang

            No harm, no foul. It was originally an offhand appendix to my main question about the Völkerwanderung, not really intended to illicit debate. I’m aware of your depth of knowledge on religious history, and I’ll be curious about your thoughts on how religion worked in pre-modern India when it comes up by and by.

          • razibkhan

            one thing that came to mind: the pala seem to have come out of nowhere. my previous hypothesis was that the ‘east asian’ component of eastern bengalis had a non-trivial munda element. i am not so sure now. rather, if it is a pulse admixture then it looks like this might be ‘overflow’ from the migration into burma of tibeto-burmans. let’s take that as a given: if you are an alien ethnicity to a region and had the choice between buddhism and puranic hinduism, what would you pick??? it seems plausible that as the less ethnically colored religion there would be a gravitation toward buddhism (working by analogy with north china and its barbarians in a similar period). therefore, a model which i think may have some merit is that the revival of ‘bengali’ buddhism may have had something to do with its role in the integration of tibeto-burmans into the population, especially at the elites.*

            unfortunately we don’t have much history to go on here.

            * a tibetan conqueror of kashmir converted to islam, after being rebuffed by brahmins who refused his entry into kshatriya status.

  2. Jatss

    still dont understand why the northern euro is so high in Haryana Jatts and even West UP samples. is there is a specific reason for it? looking at the history of the Haryana/UP area, can anyone tell something?

  3. onlooker

    A simple explanation of the jatt mixture in Punab and Haryana might be simply be the conventional one; that they refused to adjust to Brahminical caste restrictions and were downgraded. Jats in Haryana and Punjab are rather contemptuous of Brahmins and think no end of themselves as a community. Possibly they formed a mlechha group that never comfortably integrated into the caste structure.

  4. Santosh Rajan

    Jatts seem to have twice the amount of NE-Euro admixture compared to its surrounding populations. Checking HarrapaWorld Jatts also seem to have have thrice the amount of mediteranean admixture (3%), compared to its surrounding populations (0 – 1%). This is indicative of admixture with a balkan population which is high in both NE-Euro and Med components. Alexanders army after all?


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar