Hi Christian, if you don't specify any selections for occupancy refinement manually it will be done automatically just right. If you still want to do it manually, then the correct selection would be (which is equivalently to what phenix.refine does automatically in this case) refinement { refine { occupancies { constrained_group { selection = chain A and resseq 130 and altloc A selection = (chain A and resname PMS and altloc B) or (chain A and resseq 130 and altloc B) } } } } In plain English the above selections translate as: - there will be one refinable occupancy for alternative conformer A of residue 130 in chain A; AND - there will be one refinable occupancy for alternative conformer B of residue 130 in chain A, and residue PMS (*); - the sum of the two refinable occupancies above will add up to 1. (*) Since it is unique (there is no other instances of PMS in your PDB file), there is no need to specify chain id and altloc for it, so this would work too: refinement { refine { occupancies { constrained_group { selection = chain A and resseq 130 and altloc A selection = resname PMS or (chain A and resseq 130 and altloc B) } } } } Let me know if you have any questions or need more help with this! Pavel On 6/22/12 7:06 AM, Christian Roth wrote:
Dear All, I have tried an occupancy refinement for a partially modifed residue. I have split the residue in altloc A and B and the modification (PMSF) is attached on B. I included the PMS moiety as part of the residue and gave them altloc B. I created the cif link. The refinement resulted in a strange mixture of occupancys for PMS around 0.8 and both Serins 0.23 an 0.77. So PMS is not part of the residue. I am not sure about the correct syntax for the constrained group. At the moment I have used constrained_group { selection = chain A and resseq 130 and altloc A selection = chain A and resname PMS and altloc B } However the Serine pat of altloc B refines to a value not compatible to the PMS values or the Altloc A value of the serine. How do I incorporate the altloc B for the serine?
Thanks in advance for your help
Christian
_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb