Dear phenix.refine users: responding to Vijayalakshmi Janakiraman's questions about refinement at high resolution with phenix.refine, I finally documented this functionality of phenix.refine and supplied it with a small example. This paragraph (see attached HTML file) will be available as part of phenix.refine documentation in the next release of PHENIX and CCI Apps. *Important:* writing this text I assumed that the latest version of phenix.refine is used (currently, it is one from CCI Apps: /2008_01_09_2058/ , http://www.phenix-online.org/download/). Cheers, Pavel.
I have a high resolution structure with NADP as ligand bound to it. I was using CNS to start with. Somehow the Rfree was always very high. But when I tried with default options of data= and model=, in phenix.refine with the whole data and the intial rigid body refined coordinates from CNS, the Rfactors came down to 31%. I have now decided to do refinements with phenix.refine. Do you suggest any protocol for refinement of structure at this high resolution starting from initial rigid body refinements? Since I would like to handle alternate conformations in the final steps. I am wondering if the 2007 July version could handle refinements of alternate conformations. Thanks, Viji