ViewdockX Hbonds/clashes needs to be smarter about which sets of atoms should be compared

[Elaine Meng]

For induced-fit results, each ligand has its own receptor conformation, so only the interactions between those two would be relevant. Would need to use two atomspecs in the command.

There may be other pathological situations, but basically need a best guess as to which receptor conformation goes with a ligand: same receptor model for all ligands, or a different one per ligand that might be in the same submodel as the ligand.

However, it looks like the Maestro induced-fit results already have H-bonds information, so hopefully people will not be using these buttons with them anyway.