Ellipticity of thermal ellipsoids?

Kurt Thorn thorn at itsa.ucsf.edu
Wed Feb 26 02:30:44 EST 1997


I'm refining a protein structure in SHELXL-93.  I have data to 1.5A resolution,
and refining anisotropic temperature factors results in improved R(free),
but R(free) increases with subsequent refinement, so I'm overfitting my data.
I'd like to only refine anisotropic U's for those atoms that need them,
and I was going to do this by calculating how far each ellipsoid deviated
from a sphere, and restraining near-spherical ones to be isotropic.

Before I write a program to calculate the ellipticity of each ellipsoid, does
someone have code to do this already?  I'd rather not reinvent the wheel.
I was just going to diagonalize the matrix and use the major/minor axis
ration as my measure of ellipticity.

Thanks in advance.

Kurt Thorn -- thorn at itsa.ucsf.edu 

More information about the Xtal-log mailing list