Saturday, June 9, 2012

'weac2' calculator

I have made a new version of the 'weac' calculator (West Eurasian cline). This is based on a large Old World dataset at K=7 and includes the following ancestral components:
  • Palaeoafrican 
  • Atlantic_Baltic 
  • Northeast_Asian 
  • Near_East 
  • Sub_Saharan 
  • South_Asian 
  • Southeast_Asian 

The West Eurasian cline is formed between the Near_East and Atlantic_Baltic components.

Here is the table of Fst distances between components:

MDS plots of the first few dimensions:

Project participants can find their results in the spreadsheet. Non-participants can use DIYDodecad to calculate their results, but they should place all the calculator files in the same directory as the DIYDodecad software, and replace 'dv3' with 'weac2' in all the instructions of the README file.

(NOTE: Some  IDs may have wrong results in the spreadsheet because of a misalignment of IDs with results; I'll fix this and update this notice. UPDATE: Results should be correct in spreadsheet now - 9 Jun 2012)

Component labels are indicative, and you should compare your results against the normalized median results for different populations included in the spreadsheet.

Terms of Use

You are free to use 'weac2', including all downloaded files for any non-commercial purpose, as long as you attribute them to the Dodecad Project and to Dienekes Pontikos as follows:

The 'weac2' admixture calculator is courtesy of Dienekes Pontikos and was developed as part of the Dodecad Ancestry Project; more information here.


  1. Nice update, Dienekes. Thank you.
    Is weac2 one you might want to add to the gedmatch admixture tools, too?

    Also, 23andme is going to extend their chromosome painting (or at least composite admixture) analysis to include twenty "populations". Is this feasible? I estimate their database to be ~100k genomes.

  2. Actually, I'm having a hard time understanding the point of extending the weac to seven clusters, of which only two are West Eurasian. How confident are you in distinguishing between the Paleo vs West Africanness or North vs South East Asianness of people overwhelmingly Caucasian?

  3. The point is that in 'weac' populations were chosen so that the cline would emerge, and a few extra-West Eurasian controls were used, whereas now, pretty much every Eurasian population available was used, and the analysis proceeded until the West Eurasian cline emerged.

    How confident are you in distinguishing between the Paleo vs West Africanness or North vs South East Asianness of people overwhelmingly Caucasian?

    This is best done in 'byseg' or 'target' mode; if someone has strong signals for these components at individual places of their genome, then they're likely to have such ancestry, otherwise it's the expected noise.

  4. Dienekes what exactly do you mean by 'target' mode? Is target mode when you are looking at your individual chromosome percentages? For example my father and I both have some Finnish ancestry. Overall I usually score around 2.3% or so overall in this component while my father usually scores around 3%. However we both score around 12% or so in the Northeast Asian/Siberian component on chromosome 22 and we also both score around 5% in this same component on chromosome 9 when looking at components by chromosome. That seems rather high and likely to be real correct? Is this what you mean by finding real ancestry of minority components in 'target' mode?