Webhost Move

I have just moved this website over to a new domain registrar and webhost.

Hopefully, DNS has propagated and you are seeing this new location. To avoid confusion, comments on the old install are disabled.

Let me know if you run into any issues/problems. Also, is the blog faster or slower than before?

UPDATE: I had to re-import the database, so some recent comments were lost. Sorry!


  1. Looks like the site loads a little slower. Not a good thing since I moved from shared hosting to a VPS because of issues I was having.

    Anyway, I am looking into it.

  2. Hey, so we can't post comments on some of the older posts?
    Sort of OT for this post, but someone wanted to know the Admixture for the Pashtun samples from the new study.

    Zack, I hope you don't mind if I post it here. Here goes:

    Pashtun 1 : 17.5% Caucasian, 0.8% Beringian, 10.7% South Indian, 2.9% East Asian, 2 % East Asian, 3.9% SW Asian, 37% Baloch, 19.8% Northern European , 1.4 % Native American
    Pashtun 2 : 20.7% Caucasian, 12.5% South Indian, 3.7% East Asian, 1.7% East Asian, 1.9% SW Asian, 36.5% Baloch, 1.6% Med, 16.3% Northern European, 1.1% Native American
    Pashtun 3 : 10.9% Caucasian, 9.3% South Indian, 11% East Asian, 11.5% Siberian, 4% SW Asian, 32% Baloch, 3.2% Med, 14% Northern European, 0.7% Native American
    Pashtun 4 : 18.7% Caucasian, 0.7% Beringian, 12% South Indian, 3.6% East Asian, 5% Siberian, 39% Baloch, 3.6% Med, 12.7% Northern European, 0.87% Native
    Pashtun 5 : 19.7% Caucasian, 2% Beringian, 13% South Indian, 1.7% Siberian, 1.6% SW Asian, 39.8% Baloch, 14.7% Northern European, 2.4% Native American

    The main issue with my admixture analysis(which I did at k=16 as well) was that Ethiopian Jews and North African Jews decided to form their own cluster. So did the Hadza.

    Disclaimer: I don't think my run is comparable to Harappa

    My run had some HGDP Pathans at less than 10% South Indian, although majority were between 12-20%.

    • Curious, I believe your run had HGDP Pashtuns with inflated NE Euro scores too in comparison to HAP. The average was closer to 14-16% rather than the 11% or so HAP shows.

      I expect the official results on HAP to have higher South Indian, lower NE Euro and higher various East Eurasian components such as NE Asian, Siberian, Beringian, etc.

      • Yeah Paul, you are correct. I think the NE Euro scores will go down, with South Indian component going up.

  3. My bad. Looks like I can post now. Sorry for spamming.

  4. Reply to Parasar December 27, 2013

    If I may ask, why is your ancestry as per oral accounts considered to be from Kalinga? While Kalinga did have maritime connections with Lanka and SE Asia, it was not know to be a Tamil region. Was it by any chance a migration to Kalinga and back during the Chola-Ganga period?

    Good question.

    Short answer. Probably original ancestor/s were from Kalinga. As the the north became more Tamil, paternal family too would have become more tamil (or at least Tamil speaking).

    Long answer:
    Kalinga and Sri Lanka share a long history. Some of Kings of Sri Lanka were 100% Kalinga eg Nissanka Malla in the 11th century.He also claimed being descendant from the race of King Vijaya the mythical founder of the Sinhalese/Sri Lanka.

    Another telling fact is quite a few Tamils have names ending with "singham" e.g. Rajasingham, Jeyasingham and share the same names with Sinhalese, the difference being the names end in "singhe" e.g. Rajasinghe, Jeyasinghe.

    I dont think you get that type of names in Tamil Nadu, in history or now.

    • Thanks.

      At the Nissanka Malla dynasty link: Monarchs of the Kingdom of Polonnaruwa ... Chodaganga (1196–1197). Plus the consort are mentioned as: Kalinga Subadradevi, Gangavamsa Kalyanamahadevi

      This tells me that my hunch as to a Chola-Ganga connection with Kalinga is correct. It is well know that Bakhtiar Khilji conquered Magadha. What is less well known is the Tamil conquest of Magadha and Kalinga. The Tamil naval forces controlled a vast empire and managed to sail up the coast, through the Ganges and defeat the Pal emperor of Magadha from which the Pals never recovered. The Cholas adopted the moniker Gangaikonda.

      During the following period one of Chola alliances was with the Gangas of Kalinga, thus the Choda -Gangas. http://en.wikipedia.org/wiki/Anantavarman_Codaganga

  5. Just a quick question. Is anyone aware of any sort of genetic study concerning the Toda people of the Nilgiri plateau?

  6. When is zack going to post the official harappa results?

  7. @paul, would you mind posting their eurogenes k13 please?

    • I don't recall the exact results for the 2 (outliers) not used in the oracle so I'll look for them.

      However, I posted the oracle results earlier of HRP0370 and they were in response to these 3 samples. The other 2 individuals (the more South Asian and the heavily East Eurasian) were not used. Be aware that these Eurogenes K13 South Asian scores will not translate perfectly to HAP scores. The oracle showed that HRP370 is more West Asian shifted than the average of these three. Therefore, it is likely they will score higher South Indian or East Eurasian components than HRP370 did. One will definitely score much more East Eurasian. Since HRP370 scored nearly 6% East Eurasian on HAP, one of these three will likely score over 10%.

      Also, you're assuming too much from just the South Asian score. What's more important in this case is the oracle which shows that HRP0370 is slightly West Asian shifted in comparison to the average of these three. That is where you're getting that 25.37% average from. It doesn't include the more South Asian individual or the heavily East Eurasian (Tajik-like) individual.


      As for the Afghan Tajiks that IranianR1a was calling so "Iranic", their South Asian and East Eurasian scores speaks for themselves. These three Tajiks don't include the outliers either (of which one is more Pashtun like and one is even more East Eurasian).

      • The south Asian scores are pretty low for the Tajiks, especially when converting to HAP. Iranians score similar south Asian as well.

  8. Here are the results for the individuals used to compute the averages:
    North_Atlantic Baltic West_Med West_Asian East_Med Red_Sea South_Asian East_Asian Siberian Amerindian Oceanian Northeast_African Sub-Saharan

    Afghan_Hazara Hazara6_2Af 7.97 5.34 1.31 18.65 3.82 0.82 12.21 22.03 25.21 2.4 0.24 0 0
    Afghan_Hazara Hazara6_68Af 7.02 8.49 0.25 31.97 10.01 0.72 18.59 8.49 11.57 0.91 0.77 0.29 0.91
    Afghan_Hazara Hazara6_6Af 7.1 7.51 0.22 20.21 2.29 1.34 9.89 22.2 26.05 3.05 0.04 0.08 0.01
    Afghan_Pashtun Pashtun2_20Af 9.66 7.81 0.54 40.74 5.45 0.44 25.72 2.71 5.58 0.47 0.49 0.02 0.35
    Afghan_Pashtun Pashtun2_22Af 8.22 10.38 0.08 39.52 8.7 0.55 25.46 1.08 2.95 1.36 0.89 0.13 0.67
    Afghan_Pashtun Pashtun2_8Af 6.89 12.35 0.16 40.35 6.88 1.92 24.95 0.67 2.94 2.07 0.02 0.07 0.74
    Afghan_Tadjik Tadjik10_16Af 8.59 8.82 0.44 30.79 4.67 0.65 21.71 8.45 12.68 1.88 0.62 0 0.7
    Afghan_Tadjik Tadjik1_18Af 3.94 10.19 1.38 34.54 11.74 0.08 20.8 6.15 8.69 1.53 0.09 0 0.86
    Afghan_Tadjik Tadjik1_39Af 3.78 5.81 5.46 37.15 10.81 2.35 13.49 7.56 11.13 1.22 0.1 0.83 0.31
    Afghan_Turkmen Turkmen9_36Af 6.9 13.59 2.95 21.47 6.18 0.69 8.81 10.67 26.22 0.95 0.65 0.28 0.65
    Afghan_Turkmen Turkmen9_46Af 3.79 13.14 2.63 24.68 11.16 0.33 9.62 10.1 22.29 1.44 0.75 0.06 0.02
    Afghan_Turkmen Turkmen9_5Af 10.78 11.21 3.65 20.72 6.42 0.2 6.49 10.92 25.81 1.57 0.67 0.35 1.21
    Afghan_Uzbeki Uzbek8_10Af 5.89 7.69 1.2 35.9 7.59 0.84 17.53 8.76 12.68 0.97 0.37 0.23 0.35
    Afghan_Uzbeki Uzbek8_20Af 5.01 6.28 3.86 35.34 10.21 2 20.59 5.87 9.02 0.06 0.85 0.62 0.3
    Afghan_Uzbeki Uzbek8_68Af 5.15 12.71 0.37 25.41 12.96 2.13 20.28 7.6 11.36 1.44 0.51 0.05 0.04

    • Thanks for post the results HRP0282. I just linked the spreadsheet for the samples used in the oracle. I just noticed it now but all three Pashtuns used in the oracle average are more East Eurasian (East Asian + Siberian + Amerindian) than HRP0370. One is especially more East Eurasian shifted while their South Asian scores are all within 1% of one another. One is nearly 9% on Eurogenes K13.

  9. Thanks HRP0282 and Paul. So out of 5 Pashtuns, 2 are outliers...that doesn't help to be honest. I'm really looking forward to see the official results, will zack add HRP0370 as well as the kandahari pashtun to the afghan pashtun harappa results? When will zack post the results? Has he calculated them yet?

    • W, only one is a real outlier. I wouldn't call them 28% South Asian an outlier since HRP0282's Kandahari Popalzai friend (who is not in HAP) scored about 27% South Asian on Eurogenes K13. However, one is a huge outlier in that they score around 20% East Eurasian and cluster with Tajiks. One of the Tajik outliers also clusters with Pashtuns. HRP0282 and me suspect that there was a potential mislabeling of samples. The two individuals are just too distant from their own cluster and close to that of another (Pashtun vs. Tajik).

      HRP0370 is already part of the Pashtun Harappa results which include 3 Afghan Pashtuns (including HRP0370) and 1 Pakistani Pashtun. Individual samples are kept separate from ones downloaded from peer reviewed studies like Metspalu's Afghanistan study.

      I believe Zack has finished calculating the results but hasn't posted them yet.

      • So why wasn't the individual with 28% south Asian included when calculating the averages? I think your right, mislabeling could've happened. Is HRP0281 an afghan pashtun? Why hasn't zack posted yet?

        • Yes, HRP0281 is an Afghan Pashtun.

        • The individual (David) who runs Eurogenes decided not to include the 28% South Asian individual based on his own reasoning. The individual was only slightly more South Asian but his/her other components were similar to the other Afghan Pashtuns. HRP0281 is an Afghan Pashtun to my understanding but their tribal identity is unknown based on their own personal choice.

          As of now, all Afghan Pashtun samples in HAP Southern type Kandahari Pashtuns (Durrani, Popalzai, Mohammadzai, etc.) other than the unconfirmed HRP0281. The Metspalu Pashtuns are possibly Ghilzai who were banished to the North so plausibly Southern Afghan Pashtuns as well.

          For now, there have been no Northern samples from Kunar, Nangarhar, Khost, etc. In contrast, the HGDP Pakistani Pashtun are Northern Pashtun from around Kurram Valley in FATA.

          Not sure why Zack hasn't posted results yet.

  10. Iranians and Kurds tend to score 8-11% South Asian on Eurogenes K13 based on what I've seen. So, even the least South Asian Afghan Tajik scored more at 13.49% while the other two scored around 21%. However, what really differentiates the Afghan Tajiks are their 16-23% East Eurasian (East Asian + Siberian + Amerindian) scores. That makes them quite different to Pashtuns or Iranians.

  11. If zack has calculated the results, why hasn't he posted yet? Is it taking too long for him to publish the results due to the web host move?

  12. I have a completely separate question from all this discussion about pashtuns

    It's about Onge from Andaman Islands. On the Harrapa spreadsheets, it says they are about 50% south indian (component)

    Does that mean the ASI was related to the Andaman Islanders? and if so, does that make the Andamase people a south asian shifted people?

    • Interesting observation, makes me wonder if the other 50% is Oceanic/Austronesian. The Andaman Islanders are the link to the populating of Australia and the Pacific. Perhaps we should find some hints of ASI among the Melanesian islands.

      • Well the spread sheet said, they were about 51% south india(component) and yes about 30% or so related to south east asian groups like Papua New guinea and other Oceanic/Austronesian groups. I believe they even had very small amounts of West Eurasian, perhaps because of some recent mixing with outside groups. Though looking at this, it does look like they are most closes to South Asians in genetics, eventhough yes they have been Isolated for thousands of years. The ASI was probably their closes genetic cousins

  13. I cleaned up the comments. Now, please behave. I do not want any personal attacks, discussion of what happened on other blogs or forums, or useless assertions of phenotype and who is and isn't a South Asian, etc.

  14. How much African ancestry does the typical Jew have? I know they are from different regions of the world, but even many of the East European Jews, who would one think have the least African ancestry, sometimes have very curly hair and look like white skinned Ethiopians. But some others look close to Eastern Europeans.

  15. Out of some boredom I plotted the entire Harappa spreadsheet by using latitude/longitude coordinates for the various components:


    Two things I've noticed is that the Harappa component choice doesn't cluster the different ethnicities as well as some of the Eurogenes calculators. Secondly, the lack of a Western European "Atlantic" component is causing a lot of individuals to pull towards the Mediterranean when they should be pulled towards Western/Northwestern Europe.

    For some comparison, here are a few individuals with the Eurogenes calculators:

    https://mapsengine.google.com/map/edit?mid=z1DL_IP3p3uA.k4m5y2qxNrNY (K36)

    https://mapsengine.google.com/map/edit?mid=z1DL_IP3p3uA.kJKzuLSjtWHw (Old EUTest)

    https://mapsengine.google.com/map/edit?mid=z1DL_IP3p3uA.kZfduZmb6wnE (K13)

    • Thanks for posting this. I'm surprised at how it placed my data in Central Asia, specifically near Jizzakh, Uzbekistan. That's pretty accurate, given how Central Asian population dynamics are.

    • Thanks for the map, AD, but some of those results for the Harappa spreadsheet look out of whack.

      How can HRP0360 who is a Rajasthani Jatt with not more than 2% SW Asian be located closer to Shiraz than HRP0369 who is a Mumbai Parsi who plots nearer to Kerman, and who has 12% SW Asian to boot, and that who everyone knows has authentic ancestry from Iran? That looks plain wrong imo.

      Some of the coordinates for the Haryanvis who I presume are Hindus look off to me. How can these groups be located within Iran when their autosomal components are reflective of South Asian ancestry (high NE Euro and low SW Asian vs. Iranians) and they had never mixed with West Asians in the recent Muslim expansion period? You expect this for the Pakistani groups, but not for the hindu Jatts who are as caste conscious as the Brahmins, preferring only ever intermix inside their own clans.

      I think these maps should be taken very lightly until more peer-reviewed data from academia becomes available.

      • They are not geographically accurate maps, they are only relevant for relative distance between plotted individuals. Some people, like the Jatts and Europeans, are pulled sharply towards the Mediterranean (showing up as Jatts in Iran) because there is no "Atlantic" European component so Mediterranean component picks up what would have otherwise gone into that (it's really a Western/Northwestern European-like component, so they should be pulled towards the Northwest, not Mediterranean). Similarly the Northern Europeans (Scandinavians) are pulled down into continental Europe because of the high Mediterranean component in lieu of a proper Atlantic component. If you read Eurogenes blog you'll see Atlantic and Mediterranean components are related but are geographically quite far apart.

      • Btw, the Harappa "Mediterranean" component is modal in Sardinians and Basques so those are the coordinates I used (thus the southwesterly direction of pull on individuals with high "Med" numbers).

      • Almost no pakistanis have high SW Asian, stop making up stuff. Pakistanis overall have higher (Caucasian) component in them then Indians, but Caucasian is not SW Asian, it is from the Caucasus area, which is basically Southern Russia. Also the new Pashtun study shows that some pashtuns are up to 19% NE Europe component

        • sorry my last reply was for moomin

          • The study doesn't show that yet. Wait for Zack's official results. It will most likely be a few percentage points lower into the low-mid teens like the other Afghan Pashtun already in HAP.

  16. The K36 places me in Kazakhstan, so maybe the Harappa calculator is more accurate?

    • Neither are really accurate geographically, they're placing half the people in the ocean I mean, but some of the Eurogenes calculators have better ethnic clustering to space out the different groups relative to one another.

  17. These projects are meaningless unless we get thousands of samples from all the ethnic groups involved, and more crucially, they need to come from different caste groups and different regions of the state/country, considering South Asia's caste-ridden history and also the last 800 years or so Muslim occupation. Only then can you build a better profile of genetic ancestry of these places. At the moment we only have very few (sometimes less than 10 samples for a particular group) and often they might not be very reflective of the region at all. Let's take Tamil Nadu. Most of the people who have given samples are in the higher castes but we know full well the overwhelming bulk of TN is made up of lower and middle castes, who will almost certainly have far higher South Indian/lower Europoid components than the higher castes.

    • We have to start somewhere, "these projects" are in their infancy stages, just getting off the ground.

      • Well it's a start; but how many people in the lower castes will test themselves and then submit samples?

        • In India, at least some of those low caste people hold very high positions in government because of the quota system for scheduled castes and can easily afford to get some of these tests done.

          • I think the best chance will be for immigrants to the West. There are a lot of immigrants of varied South Asian backgrounds in the UK and Europe, these tests should be more heavily marketed in those markets.

          • I think this presents a problem. As per my observation, most of the people I see on these forums are looking for a scientific validating for their racial supremacy and europid affinity. Someone who is has been told and treated as socially and racially inferior would have no interest to validate the superior’s claims. Presenting a financial interest funded by a research project could be an alternative. Not sure how many people would want to spend the money solely for research and scientific self-interest.

  18. I know the national geno project did some testing and it was made into a documentary. It would be interesting to gauge a clear picture of Y and mtDNA results. The routes our ancestors took to populate South Asia is amazing. I'm surprised that hardly any N or I Y-haplogroups are found on the continent. Thoughts as to why this may be?