Dealing with indels when converting a position to chromStart and chromEnd for a bed file

Question

Starting with a chromosome # and position, I am trying to get chromStart and chromEnd values for a .bed file, but I am not sure how to calculate chromEnd when I have a variant that is an insertion or deletion.

My file is not a VCF file but it gives the chromosome number, chromosomal position (pos), reference (ref) allele, and alternate (alt) allele for each genetic variant, like you would find in a VCF file. The tricky part is that not all of my variants are single nucleotide base changes--I also have insertions and deletions that can be multiple bases long.

For SNPs I am setting chromStart = pos - 1, and chromEnd = start + 1 = pos. (Note, .bed format requires 0-based indexing and the end position is exclusive. See https://genome.ucsc.edu/FAQ/FAQformat.html)

I am unsure of how to handle indels though (insertions and deletions). Currently I am adding the length of the reference allele to the starting position to get the end position (chromEnd = start + len(ref)). I am not sure if this is right though.

I will be using this .bed file with a .bed file from Gencode and bedtools. Does anyone know the standard way the research community (or at least Gencode) treats indels when determining chromosomal position for .bed files? Thanks!

Answer 1

You can convert to VCF and use BEDOPS vcf2bed --insertions, vcf2bed --deletions and vcf2bed --snvs to segregate the VCF into three subsets.

For convenience, BEDOPS conversion tools deal with index shifts, including vcf2bed. The resulting BED file is 0-indexed, half-open.