Hi Benda,
I am refining a protein-DNA complex at around 3 A with a standard b-form dsDNA. R-factors are around 24/20 (kinda low for the resolution). Originally, I placed the DNA using coot and now, after several rounds of refinement in phenix, I get a long list of bad clashes from molprobity validation (in phenix) that covers the whole length of the DNA and almost exclusively lists hydrogens from the ribose moiety (whether or not in contact with protein). My question is,
- why is phenix.refine not taking care of these clashes (i.e. optimizes the structure such that they go away)?. Am I giving experimental contribution to high a weight (I have "optimize x-ray/stereochemistry" checked)?
because if the model is in a local minimum minimization will not kick it out of it. Simulated annealing refinement (available in phenix.refine) is the tool for this. This is illustrated on pages 67-71 here: http://www.phenix-online.org/presentations/latest/pavel_refinement_general.p... Simply make sure you are using recent Phenix version (so you get consistent clash-scores with Molprobity), add H atoms and run SA refinement. Pavel