Hello All OK back from political to scientific discussions (way more fun)!
I've trawled through the Refmac documentation and not found a way of either fixing the Biso's of a ligand (or ligands) while still refining the Bisos for the protein, or refining just the group Biso shift(s) for the ligand(s) (and the Bisos for the protein). The reason I want to do this is that I'm refining the ligand group occupancies but at medium/low resolution these are of course strongly correlated so it seems sensible to say shift the ligand Biso's so that the average is equal to the average Biso of the protein binding site atoms and refine only the group occupancy(s). As a work-around I suppose I could do it the other way around and fix the occupancy(s) at various values and refine all the Biso's, then choose the occupancy(s) which give the closest average ligand Biso to the average binding site Biso, but this seems unnecessarily laborious. I'm aware that this can be done with Buster (and probably also phenix.refine): I'm asking specifically about Refmac. Also on a related topic, a question: does Refmac take into account the alternate bulk solvent density, i.e. if the ligand occupancy is x and no alternate ordered water structure is specified, then the bulk solvent occupancy in the volume occupied by the ligand should be 1-x ? Otherwise the refined ligand occupancy(s) will obviously be over-estimated, to a degree which is probably resolution-dependent (since the higher the resolution the less will the ligand and bulk-solvent densities be correlated). Cheers -- Ian
