Thanks, Pavel, George. I think the difference is indeed marginal for the scattering function except for
the part close to f000, which may be the major reason why it works fine with the atoms. But in principle there is nothing that prevents us from using the correct Cromer-Mann coefficients for the ions by interpreting an atom record eg. Fe2+ as the appropriate ion- although our scattering functions despite being ideal seem good enough for at least standard refinement. I cannot find a figure right now but ‘ll plot a few graphs of atom vs ion scattering functions from C-Ms when I get to it… Thx, BR From: CCP4 bulletin board [mailto:[email protected]] On Behalf Of George Sheldrick Sent: Friday, June 20, 2014 9:41 AM To: [email protected] Subject: Re: [ccp4bb] refine an ion atom with different status I agree with Pavel. Even for accurate small molecule data with R-values below 3% the differences are hardly significant, the 'B-factors' compensate so well. The large majority of small molecule structures are refined with neutral atom scattering factors even if ions are present. The calculated scattering factor for an isolated ion in the gas phase is not really appropriate for the environement of an ion in the crystal anyway. George On 06/20/2014 07:18 AM, Pavel Afonine wrote: Hi Bernhard, phenix.refine makes use of charge if specified in PDB file (rightmost column after the chemical element type) to use appropriate form-factors. However, occupancies and B-factors are very efficient mops to accommodate a broad range of discrepancies between model and reality. So whether the effect of using charge is going to be noticeable, I guess, depends on the data quality (resolution, completeness, etc) and how strong the effect itself is. Also, it should be relatively easy to make a numerical experiment with calculated data to see how the total scattering brakes down into individual contributions. Pavel On Thu, Jun 19, 2014 at 6:42 PM, Bernhard Rupp <[email protected]> wrote: “.. change the valence of ion or metal except by changing the occupancy” Changing the occupancy is entirely different from changing valence. The former scales the scattering function proportionally, while the elimination of outer shell electrons predominantly reduces the very low resolution part (starting at f000) of the scattering function. Verifying the correct scattering function (e.g. Fe+++ vs Fe++ vs Fe atomic) used by the refinement program could be useful. I am curious: Garib, Pavel, Busters: How is that currently implemented? Best, BR From: CCP4 bulletin board [mailto:[email protected]] On Behalf Of Wang, Bing Sent: Thursday, June 19, 2014 10:44 PM To: [email protected] Subject: [ccp4bb] refine an ion atom with different status Hi CCP4 guys, I have a structure with heme containing an ion atom in it. Except the 4 coordinated nitrogen atoms in the heme, this ion also coordinates with one histidine residue and one ligand. But I found two negative red balls (top one and bottom one) around the ion, which is perpendicular to the heme plate and keeping in the same line with the histidine and my ligand (See the figure Ion_100 from coot in the attachment). I guess this ion has different status in it (e. g. mixture of Fe2+ and Fe3+). I simply tried the lower occupancy of ion. It clearly eliminate the negative ball at the bottom and most of the negative balls at the top, but also produced one more positive peak with slight movement instead of the negative ball at the bottom (See the figures Ion_90, Ion_85, Ion_80). The numbers in the image name represents the different occupancy ("100" means 100%, "80" means 80%). So any suggestions to solve this problem? Except changing the occupancy, is there a more precise way to change the valence of ion or metal in coot, and then refine in Refmac or Phenix? Thanks! Bing -- Prof. George M. Sheldrick FRS Dept. Structural Chemistry, University of Goettingen, Tammannstr. 4, D37077 Goettingen, Germany Tel. +49-551-39-33021 or -33068 Fax. +49-551-39-22582
