Admixture K=4, HRP0061-HRP0070

Here are their ethnic backgrounds and the results spreadsheet. Also relevant are the reference I admixture results.

The interesting samples here are the Gujarati and the Punjabi. HRP0064 is very different from the other Punjabis so far.

If you can't see the interactive bar chart above, here's a static image.

PS. This was run using Admixture version 1.04.


  1. i googled vishwakarma before you posted this. this tamil community seems to believe it was dispossessed of its brahmin heritage. from what we know so far re: genes, i am now going to put this into the "false myth" category.

  2. Since there are 14 people identifying themselves as belonging to the Brahmin castes (including the Karnataka Iyengar and excluding the 'Catholic Brahmin'), perhaps you could do a group analysis as you have done for the Iranians?

  3. Hey Zack..I just noticed my result (48) doesnt add up to 100!!..stops short @ 99.

  4. The minior components (African and E. Asian) are more intriguing to me. After all, everyone here is modeled as mostly some combination of S. Asian and European.

    HRP0027 (Carribean Indian) is listed as having 4% African ancestry. If this individual had one great-great-grandparent who was African, it would work out to 6% ancestry. Likely four generations ago (~100 years), one of this individuals great-great-grandfathers, after having emigrated to the Carribean as an indentured servant married an Afro-Carribean woman (most of the indentured migrants were male).

    It would be nice to have confidence intervals on the various fractions of ancestry so that it is possible to judge if a minor component is noise or if it is statistically significant. Does ADMIXTURE calculate this?

    • Hi, are you a Tamil Brahmin, Balaji? If so, are participating in thus project?

    • Admixture can calculate standard errors via bootstrapping. However, that takes a lot of computing time and usually I don't do it. If there is a specific set of results that you (or anyone else) is interested in knowing the standard errors for, I can calculate those but it's simply too compute-intensive to do it every time.

      • Calculating standard errors for just a few individuals should be sufficient.

        I suspect that even 1% Aftican ancestry is statistically significant. If the standard error for African ancestry in an individual such as HRP0010 is less than 0.5% we can confirm this.

        But trace amounts of E. Asian ancestry may be harder to detect (because E. Asian is not as different from S. Asian as African is). So we could start with HRP0001 who is listed as 2% E. Asian and see if the standard error for E. Asian is less than 1%. If not move on to HRP0004 who is 3% E. Asian and if that is not statistically significant on to HRP0017 (4%) and so forth.

  5. Hi Vasishta, yes and no.