Speaking of dictionaries... It seems that latest Phenix is distributed with dictionaries from 2005: /usr/local/phenix-dev-1148/chem_data/mon_lib/list % head mon_lib_list.cif global_ _lib_name mon_lib _lib_version 4.11 _lib_update 15/04/05 # ------------------------------------------------ # # --- LIST OF MONOMERS --- # data_comp_list loop_ We had a problem similar to yours with some ligands. What we did about it: get latest dictionary from Garib Murshudov's group page http://www2.mrc-lmb.cam.ac.uk/groups/murshudov/ unpack rename resulting folder monomers to mon_lib fix up the beginning of mon_lib/list/mon_lib_list.cif: Line 1: underscore missing after "global" Line 6: number sign missing overwrite the existing mon_lib folder of Phenix. It seems to work. Dmitry On 2012-09-04, at 7:55 AM, Paul Emsley wrote: On 04/09/12 12:34, Wulf Blankenfeldt wrote: this may have been asked a thousand times, but: are the ccp4 standard monomer libraries incompatible with phenix.refine, and if yes, is there an easy way to resolve this? I am refining a structure with an FMN, which I build with the "get monomer" option in coot. Unfortunately, phenix.refine doesn't like it and throws it out. Using a cif generated with elbow, on the other hand, has chirality errors (which I can fix...) and isn't liked by coot (0.6.2), which refuses to refine the ligand. :-( Hi Wulf, It seem that somehow you have an inconsistency between the model and the dictionary. The correct atom names are defined by the PDB. e.g. ftp://ftp.ebi.ac.uk/pub/databases/msd/pdbechem/files/mmcif/FMN.cif When I use "Get Monomer" I get a FMN monomer this is consistent with this (well, except for phosphate hydrogens). In particular, there are primed atom names (not starred). If you are using - or referencing an old CCP4 distribution you may be picking up an old dictionary when using "Get Monomer" - coot should tell you in the console (at least) the file names that it reads. If you want to reference the Phenix copy of the monomer library, you can use use COOT_REFMAC_LIB_DIR to point to it. If you use the model (pdb file) and dictionary cif from elbow (and call it FMN) then coot should be happy no matter what the PDB says. And I would imagine that that would be so for phenix.refine also. Paul. _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb