Separate INS and DEL events in bcftools merge results

[ahishsujay]

Hi,

I apologize if this question has already been asked, but I needed some clarification and inputs on the merge I am performing. I'm currently merging insertions and deletions for ~40 samples and while performing this I noticed instances where there are deletions happening at a particular site for a few samples but at the same time insertions happening at the same site for other samples (and vice-versa). Here's an example I found in my merged VCF:

chr10	6408	DEL79395-sample1;DEL90355-sample2;INS94796-sample3	AC	A,ACCC	.	PASS	END=6409;ChrB=chr10;StartB=11151;EndB=11151

To reiterate, there seems to be a deletion event at POS=6408 on chr10 for sample1 and sample2 while an insertion event at the same position for sample3. I was wondering if there is a way where they can be separated into separate annotations of insertion and deletion, possibly like this?

chr10	6408	DEL79395-sample1;DEL90355-sample2	AC	A	.	PASS	END=6409;ChrB=chr10;StartB=11151;EndB=11151
chr10	6408	INS94796-sample3	AC	ACCC	.	PASS	END=6409;ChrB=chr10;StartB=11151;EndB=11151

Thank you, I appreciate it!

[pd3]

Currently one can control whether SNP records can be merged with indel records (see the --merge option), but not insertions vs deletions. I will mark this as a feature request.

[pd3]

The best way to address this is to extend htslib VCF interface as follows

#define VCF_REF      0
#define VCF_SNP      1
#define VCF_MNP      2
//
#define VCF_OTHER    8
#define VCF_BND     16      // breakend
#define VCF_OVERLAP 32      // overlapping deletion, ALT=*
#define VCF_INS     64      // short insertion
#define VCF_DEL    128      // short deletion
#define VCF_INDEL  (VCF_INS|VCF_DEL)

However, as this is a breaking change in the library so may take longer