How to count variant occurrences in .vcf

Question

[darceyc17]

Hello Biostars,

I have done targeted NGS in to discover novel variants associated with a trait of interest. I am currently trying to prioritize those variants from my created .vcf to then genotype them in a larger population. Part of our variant prioritization is determining how many of the subjects, out of the 183 total, have the variant of interest. I am wondering if anyone would know how to go about this, without having to hand count each GT field, or had any suggestions.

Thank you!

[finswimmer]

Does it matter if the genotype is homozygous or heterozygous? Or is the question just "how many sample have at least one allele with this variant"?

[darceyc17]

It is how many samples have at least one allele with an individual variant, and I have 7,531 variants I need to determine this for.

[Pierre Lindenbaum]

for a start:

 bcftools view input.vcf.gz "chrxxx:12345:12345" |cut -f 10- | tr "\t" "\n" | cut -d ':' -f 1 | sort | uniq -c

Answer 1

[colindaven]

If you mean one variant and you're not planning to check your VCF any further, you might want to convert your VCF to a TSV to be able to count more easily.

This tool is available in Galaxy: NGS: VCF Manipulation VCFtoTab-delimited: Convert VCF data into TAB-delimited format

Otherwise, I like the tool vt for working with summarizing vcf statistics.

If just looking at one variant, you can probably use a BED file to specify and extract that exact variant, or even a genome browser such as IGV or JBrowse.