GATK HaplotypeCaller MNP output problem.
I have a situation here where a clear MNP is not literally emitted by the HaplotypeCaller engine.
The top VCF is where MNP distance is 0 and the bottom is 5. Looks like GATK HC engine is too INDEL centric for this kind of situations where it prefers emitting a GC deletion and an AA insertion rather than TGC - AAT substitution. I am wondering if this could be related to the pairHMM and its limitations in certain transitions (Not allowing 1 transversion and 2 transitions at the same time at the expense of introducing gaps in both sequences) types. And if that's the case I am also wondering if an alternative engine may also be implemented in GATK that does not rely on pairHMM only to detect sequence differences.
I can submit snippets of this if you would like to check.