We're sure the mosaic spread is increasing. How do we refine a slowly varying mosaic spread in Scalepack?

fit batch mosaicity 1 to 60 mosaicity will be refined separately in each batch of data. or fit crystal mosaicity 1 to 10 11 to 20 21 to 30 31 to 40 41 to 50 51 to 60 mosaicity will be refined separately for 6 groups of batches, with separate values for first, second, third, etc.. groups of 10 batches. These groupings may be different for mosaicity and for other unit cell parameters.


How do we handle a slowly varying mosaic spread in Denzo?

If you have 60 frames and you want to process first 10 with mosaicity 0.2 degree next 10 with mosaicity of 0.25 degree, and so on, up to the last 10 with mosaicity of 0.45, you may do that in the following way. 

In the command file type:

mosaicity 0.2 ... sector 1 to 10 ... go [to execute autoindexing]

@refine.dat [refine first 10 images] mosaicity 0.25 sector 11 to 20

@refine.dat [refine next 20 images] mosaicity 0.3 sector 21 to 30

@refine.dat [refine next 20 images] mosaicity 0.35 sector 31 to 40

@refine.dat [refine next 20 images] mosaicity 0.4 sector 41 to 50

@refine.dat [refine next 20 images] mosaicity 0.45 sector 51 to 60

@refine.dat [refine next 20 images] end of job

Looking at the raw data, I found that Denzo had some reflections spread over 3 or 4 frames. Given the frame size and mosaic spread, I'm puzzled by this. They were near the spindle, however, so is this a Lorentz effect, i.e., what is your criterion for excluding reflections with large Lorentz correction?

All reflections that completely cross Ewald sphere are included.


Do you only do the postrefine business on the final run after rejecting the rejects or do you do it right away? Is there an isotropic mosaicity I can refine, should I be refining in all directions, or should I really be refining just mosxx as was in the sample file somebody handed me? Does it matter as long as I'm not going to re-Denzo the frames? (I set it to .4 in Denzo but it seems to refine to around .25 or so so it seems pretty safe to me).

You can only refine isotropic mosaicity. It is better to have larger mosaicity in Denzo and refined by Scalepack, as Scalepack can reliable observed (partial) reflections as non-diffracting, but not the opposite. If mosaicity is decreased some partial reflections will be relabeled as fully recorded.


In Scalepack I can postrefine all 6 mosaicity parameters..., but in Denzo I can only use one mosaicity parameter. So how can I tell Denzo that the crystal was more mosaic in the y direction than the x? Do I change the value of crossfire?

crossfire has to do with focusing of X-rays, not reflection width in spindle angle. Both Scalepack and Denzo use isotropic mosaicity only. Anisotropic option is not yet implemented.