The question: "We have put a pyridoxal phosphate (PLP) into our torsion_information db to permit rotation about the 3 dihedral angles in the phosphate substituent of the ring. When we try tor_residue it says there is something wrong with this db. We can't see anything wrong with it. We're using O version 5.9.1. When we try a later version using tor_general, we can rotate about 2 of the bonds, but not about the one proximal to the ring. We presume the ring connectivity is preventing this. Anybody know how to get full torsion capability for a group like PLP? "
you could try to auto-generate it with MOLEMAN instead and see if that works