Use bcftools to find carriers of variants

Question

[WouterDeCoster]

I would think this is a frequent task, but it seems I cannot find a bcftools-approach to tackle this: when looking at large VCF files with (rare) variants, I often want to figure out the sample identifier of the carrier(s). I think one solution could be to add a field to the INFO field "CARRIERS=<list>", or alternatively make a tsv export of the VCF.

I now do this in python with cyvcf2, works great, but it feels like an unnecessary wrapper.

Any bcftools wizards? :)

Answer 1

[finswimmer]

bcftools query should help here, e.g.:

bcftools query -i'GT="het"' -f '%CHROM  %POS  [%SAMPLE ]\n' variants.bcf

Answer 2

[Pierre Lindenbaum]

using vcffilterjdk ( http://lindenb.github.io/jvarkit/VcfFilterJdk.html à , the following cmd will add the INFO/SAMPLES containing the name of the HET or HOM_VAR samples.

bcftools view input.vcf.gz |\
awk '/^#CHROM/ {printf("##INFO=<ID=SAMPLES,Number=.,Type=String>\n");} {print;}' |\
java -jar dist/vcffilterjdk.jar -e 'final List<String> L= variant.getGenotypes().stream().filter(G->G.isHet() || G.isHomVar()).map(G->G.getSampleName()).collect(Collectors.toList()); return new VariantContextBuilder(variant).attribute("SAMPLES",L).make();'