Dials.stills_process for ID-29 data

Beamtime Processing
1

I summarized my dials.stills_process processing pipeline for ID-29 data and presented to the SLAC Computational Crystallography Group Meeting

slides are here: ID-29 MISC Data Processing.pdf (1.1 MB)

I share some dials.stills_processes resources, show my parameter files for various dials processes, review geometry refinement strategies, and highlight various quality checks.

If anyone has any questions or feedback, please feel free to let me know!

2 Likes
2

Sweet! Big improvement in multiplicity and I/sigI! For refinement, I tend to chunk the data into 3K size chunks. Sketch of command line process like this:

  • dials.stills_process dispatch.integrate=False
  • dials.combine_experiments max_batch_size=3000
  • cctbx.xfel.filter_experiments_by_rmsd
  • dials.refine each chunk
  • cctbx.xfel.mpi_integrate each chunk

This is what the XFEL GUI does automatically. Saves a lot of time.

3

are you able to do multiple rounds of dials.refine? I like re-running stills_process and then refining again. I found I still got small improvements

4

Hi @kzielinski this looks interesting, it would be nice to catch up to get a bit more insight, we would love to add this to our portfolio :slight_smile:

5

Not automatically but I have done this.