CAPRI eval uses different atom sets for DNA in AA and CG mode

[amjjbonvin]

Module/Workflow/Library affected

caprieval

Expected behavior

CAPRI stats in case of DNA should be calculated in a consistent manner using heavy atoms (beads) of the bases (and not the backbone as it can be more flexible/variable).

Actual behavior

Currently CAPRI statistics are calculated using base atoms in case of DNA in all atoms mode, but backbone beads in CG mode, which is inconsistent.

Steps to reproduce the behavior

See the DNA_ATOM and DNA_ATOM_CG definitions in libalign.py

Suggestions on how to fix it

Replace the backbone beads by based beads for DNA_ATOM_CG

Version

1f54a9e