I'm wondering which tool or program is:

  • Most updated; doesn't use a superseded algorithm
  • Runs very quickly (ideally, under 30s per VCF file)
  • Able to be automated in some way for a large number of VCFs
  • I need to at least have these groups: Africans, East Asians, Europeans, and South Asians. More specific groups are fine.
  • Able to be run locally, as I have a lot of data

I am not looking to visualize the data or graph a PCA. My goal is to get a percentage match, or any closeness measure, for each group for each VCF file (assuming one sample per file). All VCFs are generated from microarray data.

For example, Eurogenes k13 and MDLP World-22, especially with oracle, have sufficient results but are only available as a web app.

To clarify, I am asking if there exists any such local program that exists that is peer-reviewed or validated, fast, and automate-able.

  • 1
    $\begingroup$ Solution: use github.com/stevenliuyi/admix $\endgroup$
    – BigMistake
    May 23, 2023 at 8:30
  • $\begingroup$ Please add that as an answer. $\endgroup$
    – Ram RS
    May 23, 2023 at 14:10
  • $\begingroup$ I'm not sure that it's a good solution enough to post it because as far as I can tell, they're non-validated algorithms aimed at consumers, so it doesn't address the first bullet point. $\endgroup$
    – BigMistake
    May 23, 2023 at 15:55
  • 1
    $\begingroup$ Please mention / discuss your actual problem that you are trying to solve. General "what is the best" questions don't work well for Stack Exchange questions because there's no right answer that works in all cases. $\endgroup$
    – gringer
    May 23, 2023 at 20:03
  • $\begingroup$ If you're writing it in the comments as a solution to your question, it's good enough as the start of an answer. It can always be updated / improved at a later date. $\endgroup$
    – gringer
    May 23, 2023 at 20:04


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.