*** For details on how to be removed from this list visit the ***
*** CCP4 home page http://www.ccp4.ac.uk ***
On Monday 25 December 2006 09:18, you wrote:
> A shelxl question.
> I have done anisotropic refinement of a 1.45 A structure with SHELXL.
This resolution is very marginal for full aniso refinement.
You will need strong restraints, and even so the results may not be
stable. You might want to back off to a mixed iso/aniso model,
with only the metal centers treated as anisotropic.
> I wanted to create thermal ellipsoids for an Fe-S cluster with Xtalview. 3
> of the 4 irons of this cluster are "Non positive definite" (NPD) and so
> the program uses isotropic B-factors to create the ellipsoid.
> 1.One of the iron atoms is not NPD, what could be the difference. The
> output file is from a least square refinement.
> 2. Can I manually change the anisotropic B-values of the NPD atoms to
> those of the same atoms that are not, would that be unacceptable.
That would be nonsense.
> Of course I want to do it the right way, but dont know how.
You can stop them from going non-positive definite by increasing the
strength of the ISOR restraint. But if even the most electron-dense part of
your structure are poorly-behaved, I would seriously question the validity
of aniso refinement.
For general guidance on setting restraint weights, you could have a look
at
http://skuld.bmsc.washington.edu/parvati/parvati_restraints.html
--
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742
