Exclude sites with a certain fraction of heterozygosity from vcf

Question

[thal]

Hello. I received a rather messy vcf file recently and I already did some cleaning with vcftools and bcftools. But there are two more things I want to do and I cannot find an option for it.

First, I want to exclude all sites which have more than 90% heterozygous calls. And second, I want to include only those sites which have at least one homozygous genotype for both alleles (there are only biallelic sites left). I'd be very thankful for any help.

Answer 1

[Pierre Lindenbaum]

using vcffilterjs and the following script:

function accept(v)
    {
    var nhet=0.0;
    var nhomref=0;
    var nhomvar=0;
    for(var i=0;i< v.getNSamples();++i)
        {
        var g= v.getGenotype(i);
        if(g.isHet())  nhet++;
        if(g.isHomRef())  nhomref++;
        if(g.isHomVar())  nhomvar++;
        }
    return(nhet/v.getNSamples() < 0.9 && nhomref>0 && nhomvar>0);
    }
accept(variant);

[thal]

Thanks a lot. Unfortunately I get an error.

ERROR jvarkit - Your input file has a malformed header: We never saw a header line specifying VCF version

I guess the problem is that my vcf does not has a header.

[Pierre Lindenbaum]

A VCF requires a header...