How To Remove Lines Of Certain Genotype From A Vcf File

Question

[slees.nt]

I have a VCF file containing two individuals, and I used vcftools to keep only one individual. However, the row number of the VCF file did not change, with many rows showing "0/0" genotype. I wondered if I could use vcftools to remove these rows? Below is an example of such rows:

chr10   12999368        .       T       A       4.23    .       .       GT:PL:DP:GQ     0/0:0,12,138:4:16

[Devon Ryan]

It's likely that a simple grep -vw "0/0" foo.vcf > foo.filtered.vcf would work, but you'd have to show a few lines to make sure. If nothing else, you can do this with awk easily enough.

[slees.nt]

Yes, awk should work. But vcftools seems to have GENOTYPE FILTERING OPTIONS such as "--remove-filtered-geno <string>", however, I couldn't get it work. Is there any idea if vcftools can do the job?

[Devon Ryan]

You could remove the example line with --remove-filtered-geno, but you'd also remove every other line with a missing value (.), which is probably not what you want. Just use Pierre's awk solution, it gets the job done and that's what matters.

[slees.nt]

Thanks, I agree. But I am still curious how to make "--remove-filtered-geno" work? I tried "--remove-filtered-geno 0/0", but nothing was filtered out at all.

[Devon Ryan]

When you use --remove-filtered-geno, it looks at the 7th column and removes things that match. 0/0 won't be in that column, so nothing would be changed. You might be able to --remove-filtered-geno ., but that's unlikely to be a good solution.

Answer 1

[Pierre Lindenbaum]

 awk -F '      ' '(substr($10,1,3)!="0/0")' < your.vcf

[slees.nt]

Thanks, awk is always handy. But I'm still curious if I could use vcftools to do the job.

[ivivek_ngs]

I have asked the question in the below link

To remove genotype "./." from vcf file with awk

[Pierre Lindenbaum]

ask a new question please.

[ivivek_ngs]

Thanks I will put it in a new question and remove it from here. .thanks