Slide 26 of 27
In particular, check that the scale factors are similar, especially if an absorption correction has not been performed. If there are discontinuities, then BATCH scaling is appropriate, rather than SMOOTH scaling.
- Use the ccp4i utility loggraph to check through the plots; look for smooth variations in the graphs. Unexpected discontinuities can mean trouble!
Some plots (e.g. those for Scales vs rotation range, and Analysis against Batch) should be both flat and smooth. Others will have a systematic variation, e.g. the analyses against resolution.
Several plots would be expected to be discontinuous, e.g. an axial reflection plot for a screw axis, or the plot of number rejected vs range.
If discontinuities are apparent where they are not expected, then it makes good sense to break the scaling into different runs; trying to merge batches of images together which are clearly not on the same scale is inviting trouble!