checkVCF

checkVCF copied to clipboard

According to VCF specification the "FORMAT" column is not mandatory: http://samtools.github.io/hts-specs/VCFv4.2.pdf This script fails due to line 219, that always expects a format column.

Thanks for reporting. But the specification section 1.4.2 mentions the FORMAT column should be present in the genotype fields. FORMAT can only be omitted in the header part. So I think line 219 should be kept as it is now.

I believe that if there are genotype fields present, then you must have a format column, but I think still think it is possible to have valid VCF files with no genotype columns and no format column, though admittedly this situation might be rare. As an example though, I think the vcf files containing dbSNP variants lack a format column: http://www.ncbi.nlm.nih.gov/variation/docs/human_variation_vcf/#all-00