Mid-range is hard to define, really. There are weaker electrostatic effects, and the pi-stacking and/or cation-pi effects are over longer ranges than h-bonds ty

Mid-range is hard to define, really. There are weaker electrostatic effects, and the pi-stacking and/or cation-pi effects are over longer ranges than h-bonds typically are. The whole molecule is the ligand: when you state you are "trying to distinguish which part of these molecules is the ligand," I again state that the whole thing is the ligand. The parts that bind to moieties in the active site vs. the parts that might not do so directly are equally important. In your example with amphetamine, I can assure you that the electron-rich conjugated aromatic ring is very important, even if you don't see hydrogen bonding. Remember, we are talking about overall steric and electrostatics here. If you want a bromide ion, just use a bromide salt: that directly gives you Br^{-1}. Swapping a single atom in a molecule can drastically change PK/PD properties, so caveat emptor. As for chlorhexidine, it's an antibacterial on the upper limit of breaking at least one of Lipinski's RO5, although that's not a showstopper. There is no guarantee at all that it would behave like a similar compound.