I have what seems like a deletion in my sample, that is not called by UnifiedGenotyper. The data has gone through the best practises pipeline (duplicate marking, indel-realignment, base recalibration) before calling the variants. This is what it looks like in IGV:
The mapping quality is 60 for all reads that overlap the deletion, the cigars are 100M except for the three reads that have the deletion.
I ran UnifiedGenotyper with as low as -stand_emit_conf 4.0 but there is still no record of an indel at this position, neither in the raw call set or in the call set after variant recalibration.
Why isn't this deletion called? What else can I do to shed some light on this?