*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***
If a lot of atoms go NPD (especially the Fe atoms that should be well defined) you should consider:
1. Did you refine B-values in the data integration and scaling? Maybe the values applied are unreasonable. High backgrounds, ice-rings and poor scaling of the low and high-resolution passes could also cause such problems.
2. It is possible that the quality and (real) resolution of your data do not justify full anisotropic refinement. It is the 'data to parameter ratio' that is important here, so if you have a low solvent content you need a higher resolution to refine anisotropically.
3. You may need to tighten up the restraints (reduce the esds), add ISOR restraints on all atoms and increase the distance cutoff for SIMU (default 1.7A) so that it gets applied to the Fe atoms too. Use MORE 2 (or 3) to get more information about the restraints actually applied in the .lst file (but don't then send it to the printer!).
4. In borderline cases of anisotropic refinement, we often get better results with the TLS refinement in REFMAC!
George [EMAIL PROTECTED] wrote:
*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk *** Hi all, I am refining a 4500 atom str. with 20 Fe atoms at 1.35 Angstrom in SHELXL.After adding about 400 waters, I switched on the anisotropic B factors for the whole protein ( DELU 0.01, SIMU 0.025). Rfree went down 2% and Rwork 5%,but there are a lot of Non positive definite atoms, especially most of the Fe atoms in the clusters. I tried refining the clusters only anisotropically, including the Fe atoms in the DELU instruction, but the problem stays. Is there any other way to deal with the NPD atoms? Thanks. Arti Pandey
-- Prof. George M. Sheldrick FRS Dept. Structural Chemistry, University of Goettingen, Tammannstr. 4, D37077 Goettingen, Germany Tel. +49-551-39-3021 or -3068 Fax. +49-551-39-2582
