Entering edit mode
4 months ago
RT
▴
10
Hi All,
I am trying to assign ancestry for my samples using the --supervised
option in admixture.
Here is what I did:
- I created a Plink bed with ref and study samples merged into one: 1000 WGS study samples + 2504 1000Genomes samples
- Created a ".pop" file with "-" for study samples and one of the below listed ancestry for 1000Genome samples in the same order as in ".fam" file.
- admixture -C 100 -s time --supervised -j8 --cv study_ref.bed 5
Ancestry from 1000Genomes are
AFR
AMR
EAS
EUR
SAS
admixuter_supervised_5.log:
Fst divergences between estimated populations:
Pop0 Pop1 Pop2 Pop3
Pop0
Pop1 0.076
Pop2 0.066 0.093
Pop3 0.031 0.053 0.071
Pop4 0.099 0.119 0.124 0.090
My question is how to assign the ancestry name to the output columns in the ".Q" file? Is this a sorted list of the 5 ancestry names from the ".pop" file? I can work out the column names manually from the 1000Genomes samples but that is not scalable. Appreciate your help!
Thanks, Aarthi