Serial Crystallography

For serial crystallography data analysis, the xia2.ssx pipeline is automatically run to process data from I24 SSX collections and VMXi grid scans. See these pages for further details on the autoprocessing pipelines: xia2.ssx and xia2.ssx_reduce. 

Rapid feedback is essential for understanding whether the data collection is successful, in order to make quick decisions about the next steps in the experiment. As of 2026, live per-image-analysis results for SSX are displayed at https://ssx.diamond.ac.uk/.

The web viewer shows the number of diffraction spots per image, the cell parameters from indexing each image, and a "pacman" plot showing the distribution of the spot counts over the fixed target chip. Clicking on a spot in the spots per image plot opens a basic diffraction image viewer in the browser.

The analysis is updated live during the data collection and returns results in near real time by using GPU accelerated algorithms. The spotfinding uses a GPU accelerated implementation of the DIALS spotfinding algorithm, while the indexing uses a GPU indexer developed at the Paul Scherrer Institute (Gasparotto, Piero, et al. "TORO Indexer: a PyTorch-based indexing algorithm for kilohertz serial crystallography." Applied Crystallography 57.4 (2024): 931-944.)

The indexing algorithm requires an estimated unit cell and known space group, these must be provided using the sample .yml file mechanism that is described in the xia2.ssx page.