How to convert heterozygous variants to ./. in a VCF file?

Question

[LDT]

Dear all,

I have a large VCF file from which I want to convert the heterozygous alleles to ./..

Do you have any idea how I can do that?

[iraun]

Hi! What are you trying to achieve? What do you mean by "converting heterozygous alleles to NA"? Could you simply discard them from your VCF? Could you describe your desired output?

[LDT]

I want to mask heterozygous sites and work only with the homozygotes. Thank you for the comment :)

[iraun]

I think if you just need to work with homozygous variant calls, you could simply generate a new VCF with the homozygous calls:

grep '^#' input.vcf > homozygous_variants.vcf
grep '1/1' input.vcf >> homozygous_variants.vcf

N.B: These commands only work for a single alternative allele at the variant position.

[LDT]

thats very handy, thank you so so much, but I really want to convert heterozygotes to ./. and keep them in the file. Maybe I could try a grep command? Also thank you so much for your time, means a lot learning from others

[iraun]

Sure, no worries, happy to help. So, an easy way could be with sed:

sed 's/0\/1/.\/./g' input.vcf > output.vcf

[Pierre Lindenbaum]

I have a large VCF file from which I want to convert the heterozygous alleles to NA.

NA ? that doesnt fit in the VCF specification. Could be ./. but not NA.

[LDT]

Yes you are right to ./. I wanted to write. I am changing the question asap

Answer 1

[Pierre Lindenbaum]

bcftools +setGT  in.vcf -- -t q -i 'GT="het"' -n '.'