Subset vcf on info field

Question

[gevikoj888]

Hi, I am starting with the latest VCF from ClinVar and I am trying to recover only the Pathogenic variations. To do so I did the following:

grep "CLNSIG=Pathogenic" ClinVar.vcf > body.vcf

and I then copy the header to this body.vcf and use it. However when I use it on multiple softwares I have the following error:

not a proper VCF-file.

So I don't know what I do wrong (?) but now I try to look for a way to filter variants on the INFO field within a VCF.

Thanks a lot

Answer 1

[Pierre Lindenbaum]

grep "CLNSIG=Pathogenic" ClinVar.vcf > body.vcf

with the command above, you're removing the vcf header.

You want:

grep -E '(^#|CLNSIG=Pathogenic)' ClinVar.vcf > body.vcf

or better+safer

bcftools view -i 'INFO/CLNSIG="Pathogenic"' -o body.vcf ClinVar.vcf