From hyang@rcsb.rutgers.edu Tue Mar 28 09:52:23 2006 Date: Fri, 24 Mar 2006 09:59:31 -0500 From: Huanwang Yang To: P.J.Briggs Cc: John Westbrook Subject: Re: SFCHECK and Re: Use of PROCHECK at RCSB [ Part 1, Text/PLAIN 88 lines. ] [ Unable to print this part. ] Hi Peter Thanks a lot for help. I have not got email correspondence yet, since you sent the list of bug fix. The version that I based on was v6.0.5. You can download modified code sfcheck.6.0.5.tar.gz Type make to compile. Use the command to run the program 'sfcheck -pdb file_name -sf sf_file_name' The file after -pdb is the coordinate file (either in CIF or PDB). The file after -sf is the SF file in CIF. We are interested in getting electron density properties in great details for each residue, main chain and side chain and each ligands. If you run the program, you will get an output file 'extract_map_*.mmcif which contains all the required items. The definition of the cif items can be seen from pdbx_refine_component . If the coordinate file is in mmCIF, the output should have label_alt_id, label_asym_id, label_comp_id, label_seq_id, since they are the key items in the mmcif. My modifications and extra calculations are the following: 1. export all the cif items according to the pdbx_refine_component . 2. implemented the real space R factor for residue/mainchain/sideside/ligand. 3. implemented density ratio for residue/mainchain/sideside/ligand. It is defined as the average density for the group divided by the average density for the structure. (see definition in the dictionary). This item is similar to the index by sfcheck. But for index, if one atom has zero density and other atoms have good density, the index is still 0. There is one more thing in my mind. I noticed that when there are ANISOU records in the PDB file, if the Determinant(U) =< 0, the atom will not be included for calculation. This will make R factor and correlation very bad, since only partial structure is used in calculation. I think if there is a problem with ANISOU, the atoms should be still included, but the ISO B should be used. Can you please mention this point to Aexei? Please let me know for any questions. Best regards, Huanwang P.J.Briggs wrote: Hi John Thanks for the positive response re updating PROCHECK. I'll keep you informed if anything comes of it. Regarding the SFCHECK extensions, it would be of immense help if Huanwang could provide me with modified code along with what the "base" version of SFCHECK was and what the changes are intended to do (i.e. what additional calculations are performed, what are the new or modified outputs and so forth). (I understand that these are distinct from the fixes to the broken CIF handling.) It would be also useful to know of any recent email correspondance with Alexei Vagin, so that I know what (if anything has already been discussed). Apologies if this seems like asking a lot. However I hope that it will make a more convincing argument to persuade Alexei to incorporate the changes. Thanks & best wishes Peter -- ___________________________________________________ Peter J Briggs, pjx@dl.ac.uk Tel: +44 1925 603826 CCP4, ccp4@dl.ac.uk Fax: +44 1925 603825 http://www.ccp4.ac.uk/ Daresbury Laboratory, Daresbury, Warrington WA4 4AD