Aniso scaling

JAMIER at RUBENS.ITS.UNIMELB.EDU.AU JAMIER at RUBENS.ITS.UNIMELB.EDU.AU
Tue Sep 24 03:06:01 EST 1996


I have a few queries regarding anisotropic scaling of data.
A  data set was scaled anisotropically ( infinity to 2.2, no sig cut off), and
then used in the subsequent refinement.  the results looked encouraging:
a 5% drop in Rcryst, mirrored by a 5% drop in FreeR.  However, when I looked
at the B-factors, they were significantly higher (20-30 A2) cf. to refinements prior
to aniso. scaling.
Does this difference arise from using all reflections when calculating the
anisotropic scaling?
It was suggested on the CCP4 bulletin board that anisotropic
scaling should be performed on a `pruned' data set, eg, 8 to 2.2A, 2sig
cut off.  But the resulting fob file would then be missing the excluded data.
However, I would like to refine against all the data.

Thanks
J Rossjohn
St. Vincent's Institute of Medical Research,
41 Victoria Parade, Fitzroy,
Victoria 3065
Australia
JAMIER at rubens.its.unimelb.edu.au



More information about the X-plor mailing list