How to describe an insertion in BED format file?

Question

[Zeng Jingyu]

Hello everyone,

I'm recently doing annotation on many single-sample SV vcf files from manta, but single-sample manta vcf freqently create bugs so I tried to convert manta vcf to bed format files.

Everything goes well but suddenly I'm confused by how to describe an insertion variant.

For example, a bed file:

# chrom Start   End Variant_type    Biologist_annotation    Biologist_ranking
4   71552   9371067 DEL 4p- syndrom Pathogenic
5   723194  777000  DUP     Benign
7   110276505   110478039   INS     NA
11  66287088    66288847    DEL BBS1 internal frameshift    Pathogenic
15  34735949    34806953    DEL     Benign
16  1565629 1571297 DUP IFT140 tandem duplication   Pathogenic

This example was from AnnotSV.

And for example:

Both two forms has the same start point but different end point. And from my view, the first form (the black) should be correct since the insertion can't change the position of the reference genome.

But appraently the AnnotSV do not think like I do, and I'm really not sure which form of bed file should I produce.

And for the example I drew, should it be:

# chrom Start   End Variant_type    Biologist_annotation    Biologist_ranking
1   1234    **1235**    INS NA  NA

or this?

# chrom Start   End Variant_type    Biologist_annotation    Biologist_ranking
1   1234    **1237**    INS NA  NA

Thanks for your time reading this.

Answer 1

[Alex Reynolds]

Perhaps one way is to simply isolate them to a separate file, i.e.:

vcf2bed --insertions < variants.vcf > insertions.bed

Then you don't need a custom BED3+ format.

[Zeng Jingyu]

Thanks, I'll try that.