Admixture (Ref3 K=11) HRP0131-HRP0140

Here are the admixture results using Reference 3 for Harappa participants HRP0131 to HRP0140.

You can see the participant results in a spreadsheet as well as their ethnic breakdowns and the reference population results.

Here's our bar chart and table. Remember you can click on the legend or the table headers to sort.

If the above interactive charts are not working, here's a static bar graph.

75 Comments.

  1. Punjab is a most intriguing region. The heavy bias for the European component as opposed to the SW Asian component, but with the Onge fraction perfectly within range for NW Indians is very interesting (in reference to the Haryanvi Jatt - HRP131). Another thing that has been on my mind is the unconventional genetic gradation across castes in Punjab vis-à-vis the rest of South-Asia. As small as the samples may be, the Punjabi Jatts (n=5), who are generally regarded as agriculturalists and land barons are marginally more West-Eurasian than the priestly caste of the same area; while the Satnami Harijan Sikhs are significantly more ASI than any of the Punjabi participants so far (Jatt, Rajput f.e HRP106, 107, 111, 135; Brahmin, Tarkhan/Ramgarhia). Why is there such a huge drop in the West-Eurasian admixture (and rise in the ASI admixture) when comparing the Punjabi Jatts and the Punjabi Satnamis when they are only one or two ranks below in the traditional caste system (which has always been rather vague in the Punjab as far as caste-identity is concerned)?

  2. One way of looking at it is because Punjab doesn't really have a caste system at all. It has a tribal system. Two tribes could have the same caste and be very divergent, e.g. I'll predict that even within Brahmins you'll find Mohyals will be different (more Euro/SWA likely) than other Brahmins. Also, afaik Khatris sometimes intermarry many Brahmin groups in Punjab and Jatts sometimes intermarry with some other groups as well. Rajputs and Khatris are the same formal "caste" but it makes no real difference because they don't intermarry much. Then there's the Gujjars. Anyway, caste ranking counts for less and won't influence things as much here as elsewhere.

    • Why do you think that Mohyals will be more Euro/SWA? I would think they should fall between Sindhis and Punjabis.

      • Mohyals are from NW Punjab (Rawalpindi, Lahore, Peshawar, etc) traditionally. I don't see why they would be between Punjabis and Sindhis. Also they have an oral history linking them to the Arab world (Husseini Brahmins). They tend to be endogamous and have only after partition begun marrying other groups because their traditional homeland is almost entirely in Pakistan now. My mother has lots of Mohyal family and they are quite distinct looking: fairly consistenly lighter skinned and have that Northwestern "look". Mohyals (Chhibber, Dutt, Vaid are typical last names) are technically Brahmins, but frankly it makes no difference what they are because they've historically behaved like a distinct (largely endogamous) tribe and balked at intermarriage with other Brahmins. Partition changed some of that (hence, me). Gujjars are very similar to this type of behavior. So are Rajputs. Even psychologically, Punjabis don't really think of these as layered castes but more like distinct tribes. Caste system slotting is definitely a secondary and distant idea for most of these people. This is what I was trying to clarify for AV: It actually makes little difference whether Jatts are in the same caste bracket as another group or not. He was using that notion in a Southern-Eastern style that doesn't operate equivalently in this part of the subcontinent.

        I agree with Simranjit - we need more Kambohs, Khatris (Sehgal, Sethi, etc) and so on. I say Punjabis here but suspect this is a broader thing, eg Gujjars are spread all across Rajasthan, Punjab, Kashmir and Haryana and at least politically they tend to act as one tribal confederation. These tribes do in fact operate more in a Pashtun-style organizations. This is probably true of in some measure of Haryana and west U.P. also where these same groups are found. Just my 2c.

        • I'm familiar with Mohyals as well as their stories (poems). Their tiles even today are akin to those of Bengalis and Gujarati Nagars - Datt, Sen (Lau), Bali, Baid, etc.

          Their websites have a lot of bravado but there may be some truth to them:
          Eg -
          "Many names of the Lau clan in Mohyal folklore and records closely match names from the Sena dynasty of Bengal, like Ballal Sen and Lau Sen. That, and the coinciding of the Lau clan's appearance in Punjab with the period when the Senas held territories North of Delhi, has led some historians to assert that the Laus descended from among the Senas."

          "Some scribes of Mohyal history including Russell Starcey have tried to find a tenuous relationship between Mohyal Vaids and the Baidya rulers of Bengal (990 to 1200 AD). The postulation is to be accepted with a hint of fanciful philosophy because its veracity is yet to be proved ... Some descendants of the Baidya kings migrated to north and established their states in Mandi and Suket."

          The Chhibber story looks particularly genuine:
          "After the fall of Sindh, the descendants of Raja Dahir moved to the Punjab. They were helped in their rehabilitation by the king of Delhi and established themselves in different places ... Praga Sain ... laid the foundation of Karyala ..."

          Specifics are mentioned in Chhibber poems that do not appear to have been made up:
          "Amravati ke kenare par, raja Dahar mal laryo ..." translation ... Dahar Mal fought the on the bank of the river Amravati
          The poem goes on to say that the Sultan's army fled across the river and victory was proclaimed in Gujarat.
          The battle of Nausari was fought near the river Amravati (now an overflow channel) in Gujarat. I am not fully clear on Dahar as he was supposed to been killed in Sindh as per the Arabs but the Mohyal poem does not agree with the Arab account on this aspect.
          A copper plate on the battle is transcribed here:
          http://books.google.com/books?vid=OCLC32804202&id=0bkMAAAAIAAJ&pg=PA2-IA1&lpg=PA2-IA1

          • The earliest inscription that I have come across with a form of the word Mohyal is from Ladnun, Samvat 1010 or 953AD which says "Dipadhana duhita Mohilo Gautami sutah" where Dipadhana's daughter committed sati after her husband a Mohila and son of Gautami died.

            Later, from Samvat 1162, we see mention of another Mohila - Haradatta.

            From their name types they do appear to be Brahmins. They are supposedly related to Chauhans, though the Chauhans are specifically called Brahmin (Vipra, Vatsakula).

        • Hrp019,
          I had missed the "hence, me" part in your post. I am assuming your mother's side is part Mohyal. If you have any Bali male relatives that have had Y SNP/STR tested, please let me know.

          There is a Bali (Arjun) who had posted this recently: "my tribe traces an oral history to this place. a few of them moved as recent as 1932 (to afghanistan/nwfp/kashmir/punjab), after iraq got independence from Britain. these people also went with the honorific title of 'hussaini-brahmins'" http://www.harappadna.org/2011/06/caste-is-not-ancestrally-arbitrary/#comments

          • Why specifically Bali? You know that Balis do not marry into Balis, correct? Each Mohyal subclan is exogamous (like any other Punjabi tribe actually). Almost by definition, any Mohyal is likely to have Bali male relatives somewhere. One Mohyal got identified as a 5th cousin of mine on 23andme. Family from Sialkot, and he has a Y-hap of E1b1b1a - which does point to the mideast/north africa. I will request him to send his data in to Zack. I'm not that hot on origin myths in general btw because so many seem basically manufactured and genetics is proving it (even right here on Harappa). To be honest, I don't buy the common origin concept for any given caste in the subcontinent, unless it is localized to a region. To me tribe is ethnicity in the subcontinent. Castes are an attempt to organize these tribes into a structure. Shared gotras across tribes may or may not indicate overlapping origins - the jury is out.

            BTW there is no statistical significance to the data here based on which you're drawing some of these conclusions. AFAIK there are only two Punjabi Brahmins here. And I am not even pure Punjabi or 100% verifiably pure Brahmin myself (though I may be a pure Brahmin - remember too I have 1/8 blood from a Muslim great-grandmother, who it is believed was from a "Brahmin Muslim" family). That's just the most convenient bucket Zack had for my mixed-up background. Let's look at top 10 Euro percentages for pure subcontinentals -

            HRP0063 UP Brahmin
            HRP0126 Punjabi Jatt
            HRP0006 Punjabi Jatt
            HRP0129 UP Brahmin
            HRP0008 Punjabi Jatt
            HRP0005 Punjabi Jatt
            HRP0033 Rajasthani Brahmin
            HRP0093 Punjabi Jatt HRP0006
            HRP0131 UP/Haryana Jatt

            Based on this, if anything, I would say caste is no distinction of actual genetic makeup in this factor among Northwesterners. I agree with Simranjit. We need more data for more Punjabi tribes. Also, we need a tribal orientation in looking at Punjabis because that is kinda how they breed.

          • Why Bali? - because they are supposed to be related to us. Y STR could confirm or put it to rest. We do practice clan and gotra exogamy, but in our case they are the same as everyone in the clan has the same gotra.

          • Sorry, I missed that. Who is "us"? Anyone related to Balis is related to all Mohyals. You can't just be related to one Mohyal subclan. That ain't the way it works 🙂

          • Is there a way to get your DNA tested in India? Almost all Mohyals I know are there.

          • HRP019, I reside in India and had to have it delivered to Singapore (and consequently India, then sent back) to get myself tested via 23andMe. No idea about FTDNA's Family Finder kit, though.

          • "Us" babhan - a type of brahman in eastern India (aka as Baleya Kshetra or the region of the Anava Balis). Who knows if this is all bunk but we now have some tools to confirm or discredit.

            The connection is based in part on books by Sahajanand Saraswati and T.P. Russell Stracey.

          • AV, is there an agency that does this via Singapore? If so, could you please post a link/contact for them? Or did you mean a personal arrangement of some sort? Thanks!

          • Just to make you happy Parasar, I am going to ask a Bali I know really well (not known to be related to me by blood, so that should make Zack happy too) if he's willing to get tested. Might take a few months though. It'll be yet another "let's extrapolate wildly from n=1" exercise, but what the hell. I wouldn't get my hopes high on any favorite origin myths actually turning out to be true though. Just sayin' 🙂

          • Hrp019 said;
            ".. One Mohyal got identified as a 5th cousin of mine on 23andme."

            I think a key to increasing project participants in terms of numbers is contacting relatives and informing them about the Harappa Ancestry Project. Relatives that share less than 1% of your DNA should be dissimilar enough to be included with you in the same analyses (I remember Zack mentioning this a while ago). Besides, many so called relatives on 23andMe are in fact probably not really related to you, but are turning up as an artifact of the endogamous nature of certain groups (case in point : Brahmins). I have informed a few relatives about the project, one of them submitted their data, while the others, like the majority of Indian 23andMe users, are much too inactive (and perhaps also disabled 23andMe notifications reaching their e-mail inbox) to have bothered to have gone through the message. But, it's still worth a shot, especially for the folks who belong to under-sampled populations, and populations of interest that need to be studied beyond the sparse sampling HAP has in terms of individuals to come to any logical conclusions. For me on the other hand, not too useful as not only is my ethnicity the best represented group in the project as a whole, but we are also far too homogeneous (at-least thus far) - so no surprises are likely to come up. The general population structure is much too clear, but the other participants should really go about contacting their relatives if they wish to infer some sort of pop. structure in their ethnic group - it is very likely that the predicted cousins on relative-finder belong to the same ethno-linguistic/caste group, or at least very similar.

          • Hrp019,
            Thanks. Please recommend your Bali relative the FTDNA route (or 23andme + FTDNA STR) so that STRs can be looked at just in case the person happens to be R1a1 (plain R1a1 is pretty useless due its prevalence). Which Mohyal clan does the Sialkot E1b1b1a belong to?

    • Would you have any further reading that I can look up on regarding the same subject i.e the tribal-like workings of Punjabi society as opposed to clear cut caste-divisions? When you mention tribal; I assume you must mean similar to Pashtuns rather than Indian ādivāsÄ«s. HRP086 is a Khatri, and seems similar to the other Punjabi/Sindhi participants. But the question still persists - why is the European component elevated among Jatts while the Brahmins have European scores typical for NW Indians in general despite being, assumably, more endogamous? Would this imply that the Saka admixture in Jatts is not too much of a myth after all? I have proposed this elsewhere, but I also think the elevated Northern European scores among the Jatt participants (marginally more, of course since the UP, Bihari, Punjabi and Rajasthani Brahmin[s] of HAP almost have the same amounts) is due to admixture with the Sakas. It’d also make sense for a peasant/farmer caste like the Jatts to have been historically more willing to mix with foreigners, who would otherwise be deemed as mlecchas by the traditionally upper caste Hindus. However, the Jatts themselves hold the concept of ethnic identity, gotras, clans and endogamy highly, so my skepticism remains. This trend is very, very interesting and needs to be looked into in more detail.

      • I've replied to this before , think of jatts not as a caste, but more of a tribal group like the gujjars and the pashtuns etc. Groups like the pashtuns are also peasants and farmers, does that make them more likely to mix with foreigners? I think that kind of logic is flawed. My guess is that they were remnants of saka tribes and other tribes like the white huns who mixed with the local, the other possibility is that they were an offshoot of some rig vedic tribe, like the pacytans (pashtuns). It's too early to say really.

        • Simranjit, thank you for the comment. Do note that I've considered both possibilities and have added a disclaimer at the end of my previous comment that I am still a little skeptical and confused regarding the matter. I am not concluding anything as of now, but merely speculating. I do hope that in the coming months, more Punjabi individuals get themselves genotyped, especially the comparatively under-sampled groups - more Brahmins, along with some Khatris and Gujjars will be most interesting.

          There are also a lot of unidentified (in regard to actual community/caste background) Punjabi participants; so it'd be great if they could specify the same to Zack in order to infer these results this project is giving us, better.

          Not coincidentally, Razib Khan has a post on Sepia Mutiny regarding the subject (a post fueled by the discussion here). I agree with his main point, something which I have suggested since this very debate began in the past (f.e - here) - the Punjabi-speaking Jatts indeed seem to exhibit slightly more European-specific exogenous elements than the other Punjabi castes, which might well be actual Scythian admixture, but this is only an overlay over a Vedic base. It is no myth that the Rajputs, Jatts and Gujjars have strong overlaps with regards to their gotra/clan names.

          I do wonder though - will this be restricted to certain clans? Will any other castes in the Punjab exhibit the same (elevated EU scores) as the advent of Sikhism somewhat eased up the concept of Caste in Punjab, perhaps leading to multilateral gene flow between the different communities?

          • What we need is others from the khatri , gujjar and kambojs community to participate. Perhaps some of the vanilla punjabis are. I don't think we can settle the origins issues until we can do dating of admixture.

      • "The Jatts themselves hold the concept of ethnic identity, gotras, clans and endogamy highly, so my skepticism remains" - I was confused by this. Isn't this exactly the Pashtun model? Gotra is only a mechanism to enforce exogamy afaik.

      • 23andMe delivers to Singapore (via DHL when I did so), so if you have any relatives/family friends you want it delivered to, you shouldn't have a problem. However, if you deliver it to Singapore and then have it forwarded it to India, the trip back might take a little long - customs are a little uptight about liquids and spit samples and the like as far as the trip back is concerned (i.e India to Singapore/abroad). They might even ask you to present the papers by 23andMe (such as 23andMe's permit letter to the delivery service).

        • ^ Directed at Hrp019, apologies for the error.

        • Thanks for the info AV, sounds inconvenient without someone in Singapore. I'll ask a few relatives and see if I can just take a few with me when I go next - in checked-in baggage.

          • An alternate option is to have the kit delivered to your U.S or U.K residence and then carry the kit along with you in your baggage to India, have the relative give his spit sample -> carry it back and thereon send it back to the 23andMe lab. I had it done through Singapore via a family friend (he forwarded the kit to me via DHL, again) to be on the safe side, as I didn't want to risk a relative being question about the kit by, for example, the American customs.

          • That's my intent.

  3. The Satnamis in the Ref. 3 populations are from Chhatisgarh not from Punjab. They also have significant East Asian ancestry.

  4. My own theory is that jatts are late arrival to India from Afghanistan/pakistan area after Islam took over. Prior to entering India, they were probably quasi Buddhist/hindus, and had their own customs regarding widow remarriage etc. They are agriculturist currently, and were probably farmers in their Afghanistan area. There is considerable pride in jatt heritage amongst them, so I sense that a large number of them resisted adopting a foreign religion, and decided to go south east to a more fertile land. Dr. Mcdonald places admixture of all 5 jatts in Simran's blog close to afghanistan people, and my own paternal oral history claims that we came from that area (I am HRP131) in Harrappa Project. The Scythian horseman nomad and agriculturist nature of today's jatts seem contradictory.

    • Pakhtoons from the Kandahar-Jalalabad region should be very similar to Jats as both are from Sindh. As with the Punjab north of Multan, Kandahar too was settled by Jat tribes from Sindh (the Jatan-i-Gharbi, Jatan-i-Sharqi, of Chachnama). Afghans west of the Hindu Kush will be akin to Tajiks, Uzbegs etc, though even in Bactria, especially in towns the population is more akin to the east ie Punjab & Sindh. There were also Jat tribes that lived in the Indus delta wasteland called Jatan-i-Dashti.

  5. Simranjit, to my knowledge, among the vanilla Punjabis there are three Lahori Rajputs - HRP106 and 107. HRP111 is their child. HRP073 and HRP136 are Tarkhans. No idea about the others, however.

    Vivek, your hypothesis is interesting but somewhat implausible as well - how on earth could a community that migrated to Northwest India only upon the advent of Islam further West come to be the majority caste of the Punjabi speaking areas (along with being a sizable ethnic element in northern Rajasthan, Uttar Pradesh, Himachal Pradesh, etc)?

    In light of the Saka debate, it'd definitely be interesting to see whether the Kamboj too will exhibit elevated North-Eurasian (and perhaps also SW Asian - considering they are an Iranian group) scores, since they are cognate with the Indo-Scythians.

    • Ah damn. Zack, could you possibly remove the italics from the above whenever you have the time?

    • What are Vanilia Punjabis to begin with? I am not familiar nor comfortable with that term.

      • IDK, I assumed Simranjit uses the term to denote unspecified Punjabis (nonspecific regarding caste or tribe origin, i.e).

  6. AV, Before Britishers drew boundaries people routinely use to go to afghanistan and NWFP. Essentially, over time some population that use to be centered in Pakistan/Afghanistan area move south about 150 miles over 800 years. They were probably a large group to begin with in the area, and with a component migrating south east, they became a majority group in Punjab and Haryana. You will hardly find non-muslims in afghanistan. People generally have a respect for their ancestry and will not convert out of the religion to which they were born too. My hypothesis is some jats moved south to a more fertile land and friendlier surroundings from a religious point of view. Most Jats are agriculturist. If we are to imagine a Scythian tribesman theory, its hard to imagine all of them deciding to take up farming as an occupation.

    • Vivek, who do you propose inhabited Punjab prior to the hypothetical migration of the Jatts? If they are recent migrants, why is their such a strong overlap between Rajputs/Gurjars and Jatts with regards to Gots and clan names? Also, no one is claiming that the Jatts are in their entirety descended from Sakas, but rather there has been some geneflow, albeit a minority in the Jatt genetic canvas a whole.

  7. SNORE... 🙂

  8. Let me preface my comments by saying that we certainly need more data to come up with any meaningful conclusions. That said, HRP 019 top ten subcontinental list with european mixture is wrong. Its easy to look this up. The top ten in descending order are:

    HRP 131 Haryana/UP Jatt 27%
    HRP 93 Punjabi Jatt 22%
    HRP 5 Punjabi Jatt 21%
    HRP 33 Rajasthan Brahmin 21%
    HRP 21 Kashmiri 21%
    HRP 8 Punjabi Jatt 20%
    HRP 129 UP Brahmin 20%
    HRP 85 Thathia Bhatia 19%
    HRP 6 Punjabi Jatt 19%
    HRP 126 Punjabi Jatt 18%

    Six of the ten are jatts and the Jatt with the lowest european component finds themselves on the list. I would be the first to say that sample size is too small, but there is a trend emerging at first look.

    AV, the jats probably displaced dalits in the agricultural lands who were practicing farming. I can't see Brahmins, khatri or rajputs engaging in farming in large numbers due to perceived low prestige with that occupation. As far as the shared gotra between Rajput and Jats..is a simple answer. Jats who agreed with high hierarchy of brahmins in society were rewarded with a higher pecking order in the caste system and annointed rajputs. Example of people of the same gotra name being jats in one village and rajput in an adjacent village in UP are too numerous to count.

    • There is some variation in the numbers depending on the source data.
      HRP0093=HRP0006: 22, 19

      Rajput is a ~1000 year old nomenclature, they can be of any caste or tribe (Brahman, Jat, Khatri, Hoon, Vaishya, etc). Jats are known from the period of the Arab invasion of Sindh.

    • Vivek, sorry I was only listing Brahmins and Jatts since that was what was discussed. Again, this basically shows no real caste correlation afa I can tell. You have Bhatias (Rajput/Khatri), Jatts and Brahmins all in there. If Hrp129 is here, it would be great for him/her to post locale origins. I suspect it's going to be West UP, which has an ethnic/intermarriage continuum with Haryana and Rajasthan. Disclosure: Hrp021-Kashmiri is my spouse.

      It's not true at all that Brahmins and Rajputs don't farm. Quite the opposite, actually. West UP and Haryana are a classic case: Tyagis (Brahmins), Jats, Rajputs (eg Chauhans) and Gujjars are all hard-core farmers and are all land-holders - and have been forever as far as anyone can tell. Except for their tribal affiliation, they're all kind of alike and they kind of all look the same too, actually. Their caste in practise means very little. Their tribe is what helps them compete with one another. It's interesting to see that every one of these groups has some myth about how they took-up farming, but they're all doing the same thing and in no way act like layers stratified one over another. There are other things which are unique to the Northwest. Every one of these communities has two components: Hindu and Muslim (+Sikh in Punjab) and they share surnames and gotras. Hindu and Muslim Tyagis will go to each others weddings but won't go to those of another tribe. FYI, they also claim they are Takkas, a Scythian tribe. They have limited intermarriage with Rajputs but wont marry other Brahmins. In Pakistan, there has been a slow and steady assimilation (of Muslim Tyagis) by the Ranghar (Rajput) community, with increasing inter-marriage between the to groups.

      Frankly, I don't think the classic concept of caste layering exists in the northwest. Maybe it never has or hasn't for at least a thousand years, who's to say? Mohyals are supposedly Brahmins, so why are they exclusively only mercenary soldiers? Khatris are supposedly soldiers, so why are so many of them traders and merchants? Tyagis are Brahmins, so why are they exclusively farmers and soldiers? What are Gujjars doing being large landholders? Something doesn't add up. The professions of too many groups here are at total variance with caste affiliation and they've all got reasons why. And why are they all breeding zealously within the tribe and totally balking at breeding with others of their same "caste"? It all smacks of retrofitting too much.

      If Razib were from the northwest, he would be all over this. Problem is that his Bong telescope is looking through a lot of Hindi belt haze.

      • Very true on the farming issue.

        "they also claim they are Takkas, a Scythian tribe."

        This must have been claimed for them by some Company or British bureaucrat since I really cannot imagine the Tyagis (Gaur Taga) claiming that originally. Similarly for the Jats and Rajputs, until these bureaucrats made up theories based on their whim or fancy, no Jat or Rajput was a Scythian/Shak etc.

        • This is true. A lot of the britsh clamied this or that based on whim or fancy.

          • The thing with the British was, they refused to believe that groups like the Rajputs, Jatts and Brahmins or region-wise Panjabis or Kashmiris in general were native to the subcontinent. To add to that the British colonials were fascinated by, for example; the highly sanctified environment the Brahmins lived in with constant tapas; subjecting themselves to live in optimum conditions; spiritual hygiene and cleanliness; the idea of caste identification, discrimination and caste purity. In the case of the Punjabis, it should have been their strong ethnic identity and pride along with their martial spirit. There are many mor examples, but the gist of it is all these cultural aspects greatly appealed to their somewhat racist world-view, and hence the fanciful, blatantly exogenous theories came to be.

      • "Disclosure: Hrp021-Kashmiri is my spouse.</i"

        Would you mind divulging her caste background? Or perhaps pre-Islamic caste background is she is a Muslim. Thanks.

    • Vivek, Hrp019 presented the data in ascending order, which is the default setting for the sort function in the ADMIXTURE bar chart. HRP006 and HRP0093 are the same person - the first admixture proportions are based on their 23andMe data and the second, FTDNA. I think looking at one component alone is somewhat useless. We should look at it in relation to the other components. For instance, in your case, while your European component is the most elevated for any South Asian so far (individual participant or reference populations), you also seem to have far less South-West Asian admixture than the other Jatts. In a similar vein, the Uttar Pradesh Brahmins also seem to have exogenous West-Eurasian admixture that is heavily biased towards the European component rather than the SW Asian component. For most North-West Indians though, the European admixture is slightly more than the SW Asian - both are fairly well balanced out if you take a rough view of the ratio. Either way, the European component seems to be modal in South-Asia among North West South Asians and North Indian Brahmins, which is expected.

      Re : Jatts;

      I have some reservations about this model regarding Jatts being recent immigrants to North-West South-Asia. The very fact that the Jatts are the majority caste of the linguistically Punjabi areas points to the fact that as a group (i.e farmers and barons), they have much antiquity in the area. If they did migrate later on, what did they speak prior to Punjabi? It is questionable how long the Jatt identity has been around, though. There is also much cultural and lineal overlap with other tribes of the adjacent and surrounding areas as opposed to Pashtuns. I often cite the following as the a reflective example - the Rajput and Jatts bear similar titles, such as Rao, Rawat, Harawat, Chauhan, Solanki, Parihar, Parmar, Thakurela, Chhokar, Thenua, Chapotkat, Rana, Godara, Mithe, Chatte, Khatte, Janghare, Bhagaur, Lohchab, Thakur, Antal, Malik, Gathwale, Jatrana, Chaudhari, etc. Many Jatt clans go as far as to claim direct descent from Rajputs. For example, the Ahlawat clan claims to be descended from a Chauhan Rajput of Jaipur some 30 generations ago. Similarly the Bairwal who also claim descent from Birkhman, a Chauhan Rajput. The Bacchal Jatts claim are said to be descendants of a Taoni Rajput. I don't think Jatt 'conversion' to Rajput status took place, but rather the connection between the two groups is more ancient. Farming is not in it's entirety exclusive to traditional caste Shudras per se, not even in non-North Western Indian regions. The Bhumihar Brahmins of the Hindi belt are land-owners and agriculturalists. Throughout history, it is not the Varna that has remained constant for any group, but the Jaathi. Hrp019 has illustrated this in his above comment regarding the U.P and Haryanvi Brahmins, Rajputs and Gurjars very well. An example from down south; the maternal side of my paternal lineage (i.e father's mother) were traditionally land-owning Tamil Brahmins.

      Parasar, what makes you say that the Rajput identity is not absolute? While there might be non-Rajput tribes claiming Rajput affiliation, I think their identity as lineal Kshatriyas is very clear, although through the passage of time many have taken up agriculture.

      Hrp019, I agree with you - I think clear cut caste stratification seen in North/Central and to a lesser extent South India has never existed in the Northwestern parts. Its possible that the very concept of strong caste identity, notions of caste purity and endogamy developed upon the advent of Vedic people in the Cow-belt. Why so? Perhaps the denizens of the Indus valley wdidn't significantly differ from the Sanskrit speakers biologically - it is a well known fact that the Vedic Aryans put lighter skin (disclaimer: subjective), finer features, etc on a high pedestal.

      As an aside, people have been asking about the Thathai Bhatia. While the title Bhatia is borne by Khatris, Rajputs and Jatts; the info about the Thathai Bhatia in specific seems to point out that they are Sindhi Rajputs who trace their lineage to Jaisalmer, Rajasthan. That settles that then. Interestingly enough, the Thathai Bhatia has the least Onge admixture among the South Asian participants.

      • I think we probably cannot reach any meaningful conclusion with the results at hand other than the following.

        1.) The Onge component is a clear indication of South Asianess(anything in double digits). The labelled South Asian component is probably a mix of a bunch of influences(e.g west asian)

        2.) Non brahmin south indians almost always lack european admixture , while south indian brahmins almost always have a non-trivial amount of european admixture.

        3.) The punjabis participants so far are generally very much similar irregardless of caste. This could be said also for most of north west south asians if not for the gujuratis, who seem less homogenous.

        • Agree with all your points. What surprised me about both the Gujus, bot the A and B clusters is how much they differ from Sindhis despite the geographic proximity.

          As for your second point, clearly, the population structure among South Indian Brahmins seems to be fairly clear (very homogeneous at that!). But .. one can't help but wonder whether there's a small but notable segment of variation among South Indian Brahmins that hasn't been discerned just as yet - perhaps a few individuals don't differ too much from generic South Indians, or on the other hand perhaps there are some SIBs who have genetic profiles more reminiscent of their North-West Indian counterparts? While phenotype is nothing to go by in a genetic context other than rough correlations; I am very interested in getting my paternal grandmother tested, who is rather light skinned with grey-green eyes. It'd be interesting to see whether she would be closer to NW Indians or just similar to SIBs (more likely). Anyway, I'm just thinking aloud..

          • ed, who is rather light skinned with grey-green eyes.

            just a note, pigmentation variation is controlled mostly by a few genes. in a population with a mixture of genes you'll see LOTS of intra-familial variance because the sampling error is high. this doesn't discount your thesis, but i'll be willing to give you 5:1 odds and offer $50 that this individual is not too far deviated from the rest of your family in ancestry (you'd give me $10 if i was right).

      • Re: "what makes you say that the Rajput identity is not absolute? While there might be non-Rajput tribes claiming Rajput affiliation, I think their identity as lineal Kshatriyas is very clear, although through the passage of time many have taken up agriculture."

        Rajputs indeed have been Kshatriya for a long time but have no discernable connection to the Kshatriya clans of ancient India - the Shakya, Licchavi, Madra, Gandhara, Buli etc. http://www.archive.org/stream/someksatriyatrib035136mbp/someksatriyatrib035136mbp_djvu.txt

        They appear for most part to have assumed the Kshatriya mantle about the Gupta period.

        Gohil Rajput: Rana Kumbha, 1433AD his prime ancestor Bapa Ravala, a Brahmana, 'Eklinga Mahatmya', compiled in the time of Rana Kumbha, states: Guhadatta, a Brahmana coming from Anandapur, founded the 'Guhila' dynasty.” 'Rasikapriya', a commentary by Rana Kumbha himself, on the "Gita Govinda" of Jayadeva, it is stated: “Bappa, a Brahmana, of the Vaijavapa Gotra." Samar-simha, dated V. S. 1342 (A.D. 1285) Achaleshvara temple at Abu says the same. Chitorgarh inscription dated V.S. 1331 (A.D. 1274): Bappa, a Brahmana, coming from Anandapur, worshipped the sage Harlta. Atpur inscription, dated 977 AD - "From Anandpur came he of Vipra kula Muhideosur Gohadit, from whom came the famous Gohil tribe”

        Chauhan Rajputs: Bijolia Stone Inscription calls the first Chahamana as Samanta a Vipra of the Vatsa kula.

        Bais Rajputs - their title Bais/Vaisya speaks for itself. Harsha the emperor of northern was a Vaisya.

        Bisen Raputs have numerous manuscripts indicating descent from a Brahmin Mayur.

        Parihar Rajput: Pratihara kings of Mandor descended from the Brahmana Harichandra see Jodhpur Inscription of Pratihara Bauka.

        Kadambas: Descended of Brahmin Mayur as per their early inscriptions.

        Pulakesin Chalukya – mentioned as a Kshatriya.

        As you can see, the Rajput variety is immense.

    • HRP 136 Tarkhan/Ramgarhia is at 18% as well. It would be interesting to see more Tarkhans in the project.

    • "HRP 136 Tarkhan/Ramgarhia is at 18% as well. It would be interesting to see more Tarkhans in the project." Sorry for the same comment below, I replied to the wrong post.

  9. @AV and Parasar (got tired of no threading so moving to bottom) ************

    AV you said - Besides, many so called relatives on 23andMe are in fact probably not really related to you, but are turning up as an artifact of the endogamous nature of certain groups (case in point : Brahmins). I have informed a few relatives about the project, one of them submitted their data, while the others, like the majority of Indian 23andMe users, are much too inactive

    Good God, how many relatives have you found on 23andme???? I got two "fifth cousins" and the well's run dry. One hasn't made contact with me. The other has and very likely is related to me. We haven't talked enough to pin it down completely but it is likely on the female lineage line. From what I am told, that's pretty lucky, because North Indians often (usually?) find no relative matches at all unless it is real relatives they've themselves dragged into 23andme. I have sympathy with people who do not respond or go inactive because some of this is too much family info going too public too quickly for many folks (I have misgivings myself but am getting more laidback with the whole thing).

    Parasar, I do not feel comfortable revealing any more about Sialkot E1b1b1a without his consent, except that he is not a Bali (ie is from one of the other 6 clans). BTW, for my non-blood Bali relative, can't someone get their Y STRs from the 23andme raw data? Kind of pricey to do multiple of these things (23andme and FTDNA). How many Y-STR markers make it a clincher from clan origins perspective?

    • Thanks, the negative info is useful too. The consent issue is very understandable. 23andme does not provide STR data. For Y STR 67 markers are considered a minimum to confirm. Now that we are seeing downstream SNPs, fewer markers may be sufficient. In my case I have a repeat of 13 at DYS426 which is very rare, and someone that matches me at 12 markers should be related.

      • Okay, I have a few months before I can ferry over a 23andme package to him myself. In the meantime I will see if I can't somehow use raw 23andme data to tease out Y STRs. I had done the NatGeo test years ago and at 12 Y STRs (including DYS426) got 100+ exact matches with a bunch of Russians, Germans and Poles (on y-dna). Not useful at all at that level for me.

        • Oh wait, you said they don't include STR data even in the raw file?

          • That is correct, no STRs.

            "Russians, Germans and Poles (on y-dna)" Let me guess R1a1!

            Yes that happens quite often for the first 12 markers. I have no matches though. The closest is three steps away a Pole who is R1a1a1g2 (L260 within M458 within R1a1a1) and therefore in a different group altogether.

    • Hrp019, I have 65 matches thus far and am sharing with about 18 of them. They have almost exclusively been south Indian Brahmins, but I also have one non-Indian match, an individual of Croatian descent. At least 6 of those relatives seem to be very distant, but known and confirmed via my maternal side. From what I gather, my RF matches @ 65 individuals is quite high for a South Asian - South Indian Brahmins indeed do seem to be into genotyping. Regarding Y-STR vs 23andMe - I think the latter is far more useful overall as it furnishes your autosomal DNA raw data. On the other hand, y-DNA is a very small part of your overall genome, and since doing both can be pricey, and ordering only the Y-STR kit can be compromising on, say, participating in such projects; I'd recommend 23andMe as the first port-of-call before undertaking anything else. If people are that interested in the y-DNA/mtDNA haplogroup demographics across India (Parasar does, evidently), we could probably request Zack to make a public spreadsheet, editable via a Google account wherein participants (or participants via a message to Zack?) could enter their paternal and maternal haplogroups against their alphanumeric ID. This, perhaps combined with the y-DNA results from the India DNA project over at FTDNA should shed some light regarding the y-DNA and mtDNA distribution among South-Asia's ethnic groups. This should be useful especially for mtDNA as it hasn't been studied on the basis of region or ethnicity in South-Asia (in any peer-reviewed study to date). y-DNA on the other hand seems fairly predictable.

    • Hrp019,

      Any luck finding a Bali or any Mohyal R1a1?
      Regarding - "they have an oral history linking them to the Arab world" - I wonder if we could test some for L657.

      Please see: L657+ in Arabia
      "R1a1a L342+ , L657+ Confirmed"
      http://www.familytreedna.com/public/R-Arabia/default.aspx?section=ycolorized

  10. Yeah, R1a1a. I just checked and now that number is up to 132 exact matches including some Czechs and Swedes. Totally useless afaik. Some of them say they are R1a1a1g, so that is what I must be as well, correct?

    • "Some of them say they are R1a1a1g, so that is what I must be as well, correct?"
      No - very unlikely. There has bee no M458 found in South Asia. Those matches are likely just random and useless.

  11. HRP019
    I disagree that the data above just shows random distribution, as all of the Jat participants in the project landed in the list above. Statistically, chances of that happening are low even if you limit the pool to north Indians. Sure you need more data, but it points to a possible cluster.

    On the other subject..Sure you will find Brahmins and Rajput being farmers, but in most cases they are landowners. They do not actually farm themselves in large numbers like the jats where you would be hard pressed to find an individual that can't trace his ancestry to being a farmer. However there are always different communities even within same 'caste' which have regional practices that are peculiar to them. Making any generalities about any ethnic related subjects in this part of the world is fraught with being 'called out' as there are always exceptions. In fact the only generality is that there are always exceptions. No hypothesis is going to cover every ethno-historical pecularities found in this part of the world. However, with the genetic testing we now have a tool that can be valuable in answering some of these questions. Like many have said before...we need more data to conclude anything with statistical significance.

    • Actually, that is not what's at question. It may well be that Jats have elevated Euro compared to Brahmins as a caste. The point is that you have to take a tribe-centric view. There is no such thing as one cohesive Brahmin or Kshatriya caste in NW India. To treat that as not being the case is to take a group like Jats and another group like Yadavs - group them together into "Farmer-Dairy group" and look for statistical patterns. That would result in a distribution of those people across a wide spectrum. That's what is happening with Brahmins here for sure. In other parts of India you have actual layers in a stratified hierarchy. You don't have these tribal confederations. BTW Tyagis are all farmers and are not all landholders. There are plenty of dirt poor Tyagis. Word on them is that some of them were mercenaries allied with Sher Shah Suri and fought against Babar, who post-victory later tried to coopt them with land grants - which is how some of them made it into that echelon. Their landed group stems from that period. Babar did the same thing with Rohillas (Pathans of West UP) as well in the same region. Jats, Tyagis, Rohillas, Chauhans/Rajputs and Gujjars are the farmers in West UP. Rohillas have all tended to be prosperous, but the others all have their poor-middle-rich segments.

      Actually it is interesting to see the UP/Haryana Jat has the highest Euro percentage of all and Punjabi Jatts are lower. You would expect the N Indian cline to go the other way.

      • About your last statement , whats more interesting is the consistency of the south asian. Even with the higher euro , the south asian is still in line with the rest , the other jatts have more significantly more SW asian, which is expected as you see the same case with the pashtuns as well, might be later mixture perhaps?. The dodecad number are a bit closer with it breaking up into western euro instead.

  12. Another peculiarity...Bhatia are a very odd group as some of them are considered jats, some of them are rajputs and some of them khatris.

  13. I have made a spreadsheet listing the ANI, ASI, East Asian and African components of the HRP participants with the highest amounts of ANI from Zack's K=11 table. I excluded participants with any ancestors outside the Subcontinent within the last 2 generations. You can find the table below:

    https://spreadsheets.google.com/spreadsheet/ccc?hl=en_US&key=tiDr_vcRKe91Q1NqJZjyRdw&hl=en_US#gid=0.

    All the jatts are in the top 11 and all have higher ANI than non-jatt punjabis. However we need not invoke Scythian admixture to account for this. Most of the Punjabis who have not specified their ethnicity in more detail are likely muslims and muslims tend to be more mixed or "cosmopolitan" as Razib puts it. This would account for their lower ANI. Similarly the brahmin punjabis are probably mixed with brahmins from other parts of India.

    • Much of the East-Eurasian admixture among South-Asians is actually part of the ASI component, being picked up by ADMIXTURE as such. So far only the Bengalis have non-trivial, non-noisy East Asian admixture. Either way, I don't think it's right to conclusively separate out the EA from the ASI scores at this point. I also recall Zack mentioning that he misused the correlation with the Reich et al study while computing ASI admixture (he didn't do so for ANI). At this point, the K=11's Onge component is all we have. IIRC Zack plans to get back the ANI-ASI analysis soon.

    • Do you really think 5% is much of a differnce. For example the Ramgarhia has 68 and the highest punjabi jatt is 73%. Are we really going to concluded supposed "scythian" admixture based on that. The variety of reasons based on the percentiles are numerous.

      • A 5 % difference can be non-trivial if it is observed consistently. We need more tharkans tested, my guess is that the tharkans like the other punjabi tribes share a lot more in common with each other.

        • ANI is a catch all term for all sorts of West-Eurasian admixture in South Asia, both initial (perhaps Neolithic farmers) and later exogenous elements (f.ex; Indo-Europeans i.e ANI represents both an initial admixture and all subsequent West Eurasian admixture (prehistoric and historical admixture respectively). You can't write off the possible Scythian admixture among the Jatts simply due to the fact that they have elevated ANI scores. We have to look at the components that comprise ANI..

    • Jatts are close to Sindhis. Kandahar Pakhtoons would probably be even closer. I'm on the list at #27 (HRP0003) even though located far from Sindh potentially due to the some Sindhi/Gurjara connection in the past.

  14. Harappa Participants Haplogroups | Harappa Ancestry Project - pingback on July 26, 2011 at 8:15 pm

Trackbacks and Pingbacks: