Problem

For backbones with very dense variants, we have a mechanism to gradually reduce the max number of variants in each node to reduce the complexity. In a latest run, a sample from CPTAC-LUAD, a sample failed when the max variants were reduced to 1. It is a circRNA with 2 exons (so a small one) with ~25 mutations (very dense).

Potential Solution

When a max number of variants is set, it only limits the final variant peptide calling step, but not the intermediate step which turns the variant peptide graph into cleavage graph. So we can also apply this limit to that step so the complexity could be reduced much earlier, which is the step that takes a lot of time right now.