NCS worries

["Hazes B." <bart@mrc-lmb.cam.ac.uk>]

***  For details on how to be removed from this list visit the  ***
***    CCP4 home page http://www.dl.ac.uk/CCP/CCP4/main.html    ***


If you have a 2, 3, 4, or 6-fold NCS operator then it is probably good
practice to consider the possibility that it is actually crystallographic
symmetry. A very high self rotation function peak is probably one thing to
get you worried (but what is "very high"?). One requirement for
crystallographic symmetry, unlike NCS, is that it should superimpose the
crystal lattice onto itself. I have written a program that lets you try if
this is the case. It is available over the WWW at
http://perch.cimr.cam.ac.uk/forms/isitncs_form.html and just needs the
unit cell parameters and the NCS operation as a 3x3 matrix. If the
lattices clearly do not superimpose (small deviations normally occur due
to small errors in the NCS operation and unit cell) then you have bona
fide NCS. If they do superimpose then you either are dealing with
crystallographic symmetry or some form of pseudo symmetry. I think it
would be a great idea if data scaling/merging programs would provide a
table with Rsym values for each independent symmetry operation (and one
listing the statistics for just the Friedel mates). That might help people
to spot pseudo symmetry or other anomalies at an earlier state.
If it looks like you have NCS then it is also well worthwhile to calculate
a "self-Patterson" using your observed intensities and all phases set to
zero. If you have (pseudo) translational symmetry (this happens for
instance when a NCS symmetry is parallel but not coincident with a
crystallographic symmetry axis of the same order) it should give a
distinct peak in this map.

Bart