Allow different thresholds for insertions and SNVs

[bwlang]

In omicron SC2 sequences there is difficult area (3 bp deletion - 8 bp - 9 bp insertion). Typical aligners produce more soft clips than insertions.

Since insertions are more difficult to align correctly it would be helpful to have different thresholds for insertions vs. SNVs.

other ideas to improve calling in such regions:

• look for frequency of soft clips in this region, more soft clips = indirect evidence of indel
• indel lengths that are multiples of 3 are less likely to be artifactual in viral sequencing, maybe consider this?

[danrlu]

Is this similar to #79? May be worth consolidating so the follower of that thread can see this. It would be really helpful to have this feature in. Thank you Brad!

[bwlang]

I think it's only a bit related to #79 , if I understand it correctly that one is referring to a situation that calls an indel using an adjacent SNVs frequency.