Can anyone suggest to me how to relax a protein with a non-covalent ligand? I want to make sure the ligand-binding site does not change dramatically. In the following scenarios, which one looks reasonable:
1) relax protein with ligand; use movemap to exclude the rotamer of ligand residue
2) relax protein without ligand; use movemap to exclude the rotamer of residues contacting with the ligand. In this case, does the following movemap correct? (I only find movemap example to exclude consecutive residues or single residues rather than some random residue)
RESIDUE * BBCHI
RESIDUE 37 NO
RESIDUE 72 NO
RESIDUE 114 NO
RESIDUE 348 NO
RESIDUE 325 NO
Your movemap idea is a good one but will not work in the context of torsion space relax; movemaps control torsion space not 3d space (so being "near the ligand" isn't the descriptor you need it to be).
I would start out with just the flags
for constrained relax to start with and see if you are happy with the results. (the latter two of these flags are for constraints; the first is a scorefunction tweak that improves relax, or used to anyway)
Hi Thanks for your suggestion. Do I need to include ligand when applying your method?
Include the ligand in what? Your PDB file? Yes, if you want it to remain in the model.