How to get only polymorphic genotypes from combined vcf file filtering out homozygous, heterozygous and missing genotypes

Question

[analyst]

Hi everyone,

I have a combined vcf file of multiple samples. I have to remove homozygous, heterozygous and missing genotypes and keep only polymorphic genotypes.

Kindly suggest which tool should i use to get desired output.

Thanks a lot!

[dthorbur]

You can parse the genotype (GT) field in the vcf file, see docs for vcf 4.1 here.

It would be as easy as only taking lines more than 3 different numbers appear in the GT field. You can easily do this in R with the vcfR package. There is also this post from a few years ago that offers other options including one from bcftools, among others.

[analyst]

Thanks dthorbor!

I think this command from the post only removes homozygous alleles not heterozygous alleles. What else should I add to remove heterozygous and missing alleles as well in below command

bcftools view -e 'COUNT(GT="AA")=N_SAMPLES || COUNT(GT="RR")=N_SAMPLES' input.vcf

Thanks a lot!

[dthorbur]

I'm unsure if there are any tools that do exactly as you want as your use case sounds fairly niche. As stated, it would be easy to parse a VCF if you have a basic grasp of the genotype field and can use R (or the PyVCF package if you use python).

[analyst]

In the previous post, there is another command that keeps sites with at least one nonref allele. Can I increase stringency of the criteria to keep sites with at least 40% polymorphic alleles by replacing 1 with 40?

bcftools view -c 1 input.vcf.gz -o output.vcf.gz -Oz