Hi Ivan, I'm not sure which version of phenix you are using, but in 1.7.3-928 GUI restraints are generated for metals using the "ReadySet". ReadySet creates a file called your_pdb_file_name.metal.edits. The file I generated gave standard deviations to bond lengths of 0.05-0.25.
Prior to generating this file & including it in refinements I found phenix treated the Mg ions in the structure I was refining like waters & thus it kept "blowing-up" the coordinated ligands and waters around the Mg. I'm not sure if you've been struggling with a similar problem. There is probably a non-GUI way of running ReadySet (phenix.ready_set - maybe?). I was unable to find out from my log files where the "ideal" values in the .metal.edits file come from but they look consistent with those reported in Harding, Small revisions to predicted distances around metal sites in proteins, Acta Cryst. (2006). D62, 678–682. Genevieve Dr Genevieve Evans ------------------------ Laboratory of Structural Biology & Maurice Wilkins Centre School of Biological Sciences University of Auckland ________________________________________ From: CCP4 bulletin board [[email protected]] on behalf of Ivan Shabalin [[email protected]] Sent: Friday, 25 May 2012 5:58 a.m. To: [email protected] Subject: Re: [ccp4bb] ligand geometry evaluation Dear all, I found that thread in my search of what ideal values for metal Phenix is using. I tried to use that command "phenix.pdb_interpretation model.pdb write_geo_files=true" for my pdb file, and I get all the values for most atoms, like: >>bond pdb=" CB LEU B 332 " segid="B " pdb=" CG LEU B 332 " segid="B " ideal model delta sigma weight residual 1.530 1.511 0.019 2.00e-02 2.50e+03 9.40e-01 >> but for Mg ions i did not find the ideal value: >> nonbonded pdb=" O HOH C 405 " pdb="MG MG M 1 " model vdw 1.906 3.250 >> Can I specify some keywords to get it also for Mg? I looked through phenix folders and I found energy_lib.cif file. There are the following records for Mg: MG NPA metal 500.000 2.090 . MG NPB metal 500.000 2.090 . MG OHA metal 500.000 2.150 . # ? MG OHB metal 500.000 2.150 . # ? MG OHC metal 500.000 2.150 . # ? MG O metal 500.000 2.180 . # ? Do I understand correctly, that these are actually the default values? If I use Phenix with default parameters, then distances will tend to become ideal? Especially at low resolution when X-ray weight is low? And what does mean, that for Mg there is no standard deviation of the bonds (. # ? is placed instead)? How phenix will handle it? will it assign some default esd? I appreciate any comments! Thanks a lot!! With best regards, Ivan Shabalin, Ph.D. Research Associate at Minor Lab, University of Virginia 4-224 Jordan Hall, 1340 Jefferson Park Ave. Charlottesville, VA 22908
