Generally, can I safely use one variant caller e.g. Mutect2 if all my samples are normal matching? Because when 2 to 3 callers were used, the number of mutations decrease dramatically after intersecting the mutations.

I am looking at the somatic mutations in normal tissues and the mutant clones can be very small. Mutations that are called by 2 or more caller are mostly likely to be true variants. But my worry is missing some important mutations...

Is there solution to this?

Thanks