[ccp4bb]: MOSFLM warnings

["Leslie A." <andrew@mrc-lmb.cam.ac.uk>]

***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***


Dear Ricardo,

       The first thing to check is whether you really have the
reported slippage in the series of images that you have integrated. If
you were integrating the images in a background job, then the message
is certainly true, but if you have done the integration interactively
using the GUI, and you have also done other tasks in the same session
(eg refining the cell, integrating a few images) then sometimes this
warning is incorrectly given because the program is actually checking
between different runs (I think this has been fixed in the development
version soon to be released).

So, look at the summary file for the integration run. You can either
do this with loggraph or looking at the file itself (loggraph is
easier). You should be able to locate the image(s) where the missets
change by 0.13 degrees. If you can't find it, then there is no need to
worry.

One situation which can give rise to apparent large changes in the
crystal orientation is if the diffraction is very weak, so there are
only strong spots near the centre of the image. In these
circumstances, the orientation of the crystal around the X-ray beam
direction (PSIX) which is derived from the spot positions, is poorly
determined. Normally this is also associated with large variations in
the TILT and TWIST parameters (eg shifts of more than 10 between
adjacent images) and ROFF and TOFF for radial readout image plate
scanners. In these circumstances it is best to fix these parameterS:

REFINEMENT FIX TILT TWIST (ROFF TOFF)

If their true values (either the average from your data, or the values
from a strong dataset) are known, then input these using the
DISTORTION keyword.

If the images are not weak and you really have shifts in orientation,
you should double check that your goniometer heads do not have any
play and that the magneting coupling for your cryo-pins are clean and
strong.

Another possible cause is due to shutter problems. The crystal
orientation is refined using "post-refinement" procedures, which
examine how the total intensity for a partially recorded reflection is
distributed over several images. If the shutter is sticking slightly,
or if the incident beam intensity is not constant, this will cause
errors in the orientation refinement. If the problem is serious, it
should also show up as partially recorded reflections having
significantly worse merging statistics than fully recorded, and
probably in the partial bias statistics as well.

If the apparent change in orientation is due to weak images, you will
probably not be able to pick this up in the SCALA statistics. If it is
a genuine change for just one (or a few) images, then in should show
up in higher Rfactors for the images where the shift occurred (and
possibly the ones on either side). In this case you could consider
rejecting those images if the Rfactors are much higher.


  I hope this helps you tackle the problem,

     Andrew Leslie