Extract samples with alternative genotype from a multi-sample VCF

Question

[tacrolimus]

Dear all,

I have a multisample vcf and want to extract the IDs of those with rare impactful SNPs with at least one alternative allele at the a loci within the region of interest.

I am running the following on a normalised multi-sample vcf which has been annotated with VEP:

bcftools view -r REGION OF INTEREST | filter_vep -filter 'IMPACT is MODERATE or IMPACT is HIGH' |  filter_vep -filter MAX_AF <0.01 or not MAX_AF'' | bcftools view -i 'GT="alt"' | bcftools query -f "'[%CHROM:%PO %SAMPLE %GT\n]' >output.txt

I work in an airlocked environment with patient data so can't share the VCF I am using but it is a standard multi-sample vcf created by merging using bcftools.

I get an output with a variant and ID per line irrespective of the genotype e.g.:

chr1:123:456 sampleID123 ./.
chr1:123:456 sampleID234 ./.
chr1:123:456 sampleID156 0/1.
chr1:123:456 sampleID123 ./.

etc

I would like to just get the sample IDs with at least one alt genotype.

Many thanks

[Pierre Lindenbaum]

chr1:123:456 sampleID156 0/1.

I don't understand your problem. In your example above , there is a "sample IDs with at least one alt genotype."

[tacrolimus]

yes the 0/1 variant is the one I want but the output also includes those with two ./. variants. I only want those with at least 0/1.

Answer 1

[abascalfederico]

If you just want those lines with a "1" in the third column, you can add an awk command:

bcftools view -r REGION OF INTEREST | filter_vep -filter 'IMPACT is MODERATE or IMPACT is HIGH' |  filter_vep -filter MAX_AF <0.01 or not MAX_AF'' | bcftools view -i 'GT="alt"' | bcftools query -f "'[%CHROM:%PO %SAMPLE %GT\n]' | awk '$3 ~ /1/ { print }' >output.txt

[tacrolimus]

True, I could also just grep '0/1' '1/1' etc but I was wondering what the error with the code above was? Thanks for taking the time to answer.