Yerko Escalona wrote: > I have a problem when i model the structure of proteins with ligands > (hetatm = NADP, FAD and FMN), because some residues have improper bond, > like the Asparagine make bond between the N and C(gamma), or > Phenyalanine with wrong position of aromatic carbon.
When you build models with ligands, Modeller adds protein-ligand restraints to keep the ligands in approximately the correct binding conformation. See http://salilab.org/modeller/9v7/manual/node65.html
It is possible that these restraints are hard to fulfill in your model (perhaps because the sequence identity in the binding site is low). To attempt to rectify this, you could build without ligands (then use another program to dock the ligands into your model) or possibly alter the nonstd_restraints method or the alignment to relax some of these restraints.
Ben Webb, Modeller Caretaker