Jason, While phenix.refine is an incredible powerful tool for structure refinement, as Nat said earlier some re-building may be required. Without re-building addition of ramachandran restraints may only mask problems rather than fix them. Fortunately, Cbeta deviations are quite sensitive and as far as I know are truly independent structure validation measures see: http://www.ncbi.nlm.nih.gov/pubmed/12557186 quoting from the abstract: Cbeta deviation is sensitive to incompatibilities between sidechain and backbone caused by misfit conformations or inappropriate refinement restraints. Good Luck, -bob On Wed, Apr 6, 2011 at 9:50 AM, Jason wrote:

Hi, I have tried to regularize my protein structure in coot and use ramachandran restrain to improve statistics (and then release them). However, I can not make the c-beta deviation be zero. Is there some special way to improve c-beta deviation? Thanks.

====================== Jason Structural Biology Department University of Pittsburgh ======================

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

Hi Jason, C-beta deviations are often coupled with rotamer outliers, or even an incorrect rotamer. One technique that Paul Emsley showed me for correcting C-beta deviations in Coot is to mutate the offending residue to ALA, then regularize and real-space refine the region (including 2-3 residues on either side). This will hopefully allow for the mainchain to find a more suitable position, without a misplaced sidechain trapping it in a false minima. After that, you can mutate the residue back to what it should be, and look for the best rotamer given the density and the new mainchain orientation. Jeff On Wed, Apr 6, 2011 at 6:50 AM, Jason wrote:

Hi, I have tried to regularize my protein structure in coot and use ramachandran restrain to improve statistics (and then release them). However, I can not make the c-beta deviation be zero. Is there some special way to improve c-beta deviation? Thanks.

====================== Jason Structural Biology Department University of Pittsburgh ======================

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb