First you must be on the same origin
There are various ways to get there - brute force method is to calculate
phases from both models, combine them, and use the reflection utility
phase match to move one to the other by an appropriate origin shift...
Then
Use the GUI map correlation.
It will give you the map correlation residue for residue ( or section by
section..)..
But remember to move the second set of coordinates by the same origin
shift applied to the phases..
E
Vellieux Frederic wrote:
Dear all,
I am writing because I have a question that concerns coot or ccp4 (don't
know which one). I have 2 maps and one model. One of the map is an
experimental map that has not seen one inch of a model, the second map
is generated after refinement of a molecular replacement model. Same
molecule, obviously. They both look good. I am trying to find out which
of the 2 maps is the most informative. Hence I wish to compute a global
(real space) correlation coefficient "a la Alwyn Jones". In fact 2 sets
of numbers per map : an average real space correlation coefficient, and
a residue-per-residue correlation coefficient. The average real space
correlation coefficient I can easily compute if I have the
residue-by-residue real space correlation coefficients by writing a
small jiffy program. I'm sure that someone must have done this. I
suppose I could use CNS, but it's a bit tedious with having to generate
the mtf file... Hence I am looking for something simpler. I did not see
anything in the Phenix GUI.
Any way of doing what I want in coot? Or in ccp4i?
Thanks,
Fred.
