Investigating the scoring function in paired end mode

[ksahlin]

Strobealigns mapping-only accuracy in pared-end mode is worse than in single-end mode compared to minimap2 (See fig 2 in https://www.biorxiv.org/content/10.1101/2024.10.29.620855v1). Specifically, we produce higher accuracy mappings for short SE reads than minimap2 but worse accuracy in pared-end mode.

One would think that if individual scoring is better, it would allow paired-end scoring to also be better.

This merits investigation.