I wonder if Indelocator will provide extra indel variants that mutect2 could not call.
Good morning @bigbadbo,
Because Indelocator is not one of our tools, I am unfamiliar with it except for the fact that it was used at the start of TCGA and that it is much like a pileup caller without any statistical modeling nor reassembly. The Broad CGA currently uses Mutect2 to call indels. I think the question you are asking is how does the sensitivity and precision compare between the two tools for indel calls. This would be like comparing apples to oranges as Mutect2 performs reassembly and statistical modeling.
As with any collection of tools, the Venn diagram of overlap will always show mutually exclusive calls. If sensitivity is the aim, then I believe folks cover their bases by taking the union of callsets. If precision is the aim, then folks take the calls that at least two callers coincide with, given the groupings of tools are comparable in performance.
I would recommend you use GATK4 Mutect2 to call your indels.