Questions, issues, rational decision making, and time saving tips while building, refining, and validating X-ray structures that have alternate confirmers. I am currently trying to build/refine a 1.2 A resolution structure using Coot and Refmac5. There is clear interpretable Fo-Fc difference density in several helices/turns/strands, and I am building alternate confirmers into these areas. Several issues arise in this process.
1) Trying to build using sound stereo chemical constraints for the alternate confirmer in Coot is laborious (for me, at least) because of connectivity issues at both ends of the alternate confirmer chain back to the single main chain confirmer, as well as Regularization refinement atom selection in Coot between alternate confirmers. It seems that to build 5-6 residues in an alternate confirmation I need to create an alternate confirmer residue chain length of about 10-13 residues, about 2-3 of additional residues on either side of the alternate confirmer residues in strong difference density, to remove much of the stereochemical bond problems associated with joining confirmer B back with confirmer A. It is difficult select the entire alternate confirmer chain in Coot for restrain refinement because the two chains overlap so well at the ends of the alternate confirmer. a. To Paul Emsley and the Coot designers: is it possible to create a display chain and display alternate confirmer toggle switch in the display window manager? (I hope I am not missing a feature in Coot, or raising a question brought up previously.) 2) I just put my structure through the Molprob, and the PDB validation servers, and neither server annotated sterochemical anomalies in the alternate confirmers, which I know are there. a. How well should these alternate confirmers conform to sound stereochemical constraints, when joining up with the primary chain? b. Is there a way to validate my PDB file while leaving my entire PDB file intact, or do I need to separate the confirmers A and B into separate PDB coordinate files and submit each separately though PDB validation software. 3) If anyone has comments on how they go about this process in general (of building, refining, and validating alternate confirmers), I would be glad to read them. Thanks in advance. Have a great weekend -- David Shultis Ph.D. Molecular Physiology University of Virginia 480 Ray C Hunt Dr. Snyder Bldg Rm 360A Charlottesville, Va 22908 434-243-2730
