Entering edit mode
3.3 years ago
Alexandra
▴
20
Hi all!
I'm having problems with .evec file when running smartPCA. This file doesn't consider all the individuals that I introduced in .ped file (529 individuals).
Anyone knows what the problem might be? The command that I used was:
./smartpca -p Patag_sud_maf_ld.par > Patag_sud_maf_ld.out
The .par file contains:
genotypename: Patag_sud_maf_ld.ped
snpname: Patag_sud_maf_ld.map
indivname: Patag_sud_maf_ld.ped
evecoutname: Patag_sud_maf_ld.evec
evaloutname: Patag_sud_maf_ld.eval
numoutlieriter: 0
I will appreciate your answer!
Hey alexandra - super hard to answer without having access to the files. If I were faced with this problem, I would first check that all the sample IDs match exactly across the files. I would then check that all the files are formatted correctly.
If both of those don't turn up anything useful, my third step would probably be to find which of the 529 samples are PCAed, and which of the 529 are not. Once you have lists of who didnt get included, you might be able to work backwards and figure out why they weren't... (i.e. if those not included share any common feature).
Hope that helps.
Thank you! But still I don't know what the problem might be. I tried what you told me, and the .evec file only consider a group of samples that belong to the same region and were taken together. Maybe is the problem in the merge of the data?
Yeah that sounds like it is almost certainly the problem. I am not sure exactly what you mean by "same region and taken together" but that sounds like the kind of thing that will trip the PCA algorithm up...