Hi Ralf, Thanks for both of your replies. The case with the UNK atoms is that I have an 18 a.a. break for a loop that runs from south to north up the back side of my protein. About midway there is density that clearly is not water (I have a 3 proline stretch and I would guess that it is one of them) so I wanted to model atoms into the density. So in that case I would expect the UNK atoms to behave as if they were C, N, or O. This is the first time I've had to deal with a synthetic ligand in a structure so I generated the .cif file using REFMAC. According to Pavel's response about the UNK atoms I'll need to use elbow.builder to generate a library file that phenix.refine can use so I guess I'll just wait and see if elbow's .cif file "fixes" the problem with the ligand as well. If not, then I'll follow with your suggestion about editing the .cif file for that torsion restraint. I appreciate the help, -Jeff **************************************** Jeff Habel PhD Scripps Florida Drug Discovery 5353 Parkside Dr, RF-1 Rm 102 Jupiter, FL 33458 Phone: 561-799-8849 Fax: 561-799-8957 Email: [email protected] ***************************************** -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Ralf W. Grosse-Kunstleve Sent: Thursday, August 30, 2007 2:46 PM To: [email protected] Subject: Re: [phenixbb] ligand restraints Hi Jeff, how do you generate the restraints for the ligand? Could it be that the restraints dictate to co-planarity? In that case, could you try manually removing the unwanted dihedral restraint from the .cif file? Ralf _______________________________________________ phenixbb mailing list [email protected] http://www.phenix-online.org/mailman/listinfo/phenixbb