merge VCF single variants G>A and C>A into GC>AA

Question

[cocchi.e89]

I am working with some WES human data, I am trying to replicate some results from previous study from which I have the HGVS.c variants annotation and I need to get the genomic coordinates. I am doing pretty good with VEP but I found a couple of variants that are delins of 2 bases and resulted in splitted variants in my VCF file. As example:

NM_000179.2 MSH6 c.866_867delGCinsAA (Chr2: 48025988 - 48025989) p.Gly289Glu (clinvar page) is absent from my VCF but I have both:

2   48025988    rs368318845 G   A   103592  PASS    0.0002592   1   A
2   48025989    rs267608047 C   A   97356.5 PASS    0.0002393   1   A

that combined would result in the delGCinsAA above.

I am asking you, is there any simple method to "combine" these variants in order to know which patients are GC>AA hom, het and homref?

Thank you very much in advance for any help!

Answer 1

[GokalpC]

If they are in phase then ReadBackedPhasing will set them together but that is an old tool from GATK3 era.

Another useful strategy is to enable MNP calling with your variant caller. Freebayes does this by default or GATK4 HC can be adjusted to call MNPs with given distance.

Other than the read information that shows the phasing between alleles there is no simple tool to merge them together.