BioSAXS data collected at the BM29 beamline are processed and analysed by DAHU from sample-buffer subtration on.

Steps in autoprocessing are:

  • collected 2D images.edf are stored in .h5 file directory together with their metadata .json files (created by BsxCuBE);
  • 2D images are converted (radial integration, normalisation to beamstop diode reading and exposure time) into 1D curves applying parameters from .json files and save in /1d directory using pyFAI;
  • average 1d curve is calculated from frames of the same run, curves with some difference (parameter defined in BsxCuBE, ex. -log(fidelity) < 2) from the first one are discarded from averiging. This approach is called radation damage check, however works for any sample or beam inhomogeneities/instabilities and can be disabled (un-tick in BsxCuBE);
  • AutoRg, AutoGnome, dammif routines are used for further autoprocessing.



Python scripts for re-processing can be found on stanza, slavia and smaug

Some tips for HPLC runs:

- if HPLC run didn't appear in ISPyB (appears normally only after run, not at beginning as in SC mode) check .xml of last frame for part linked to ISPyB, if missing complete lines by copy-paste from any .xml from a run which "worked" and run python /users/opd29/workspace/edna/bioSaxsv1/bin/ *.xml in raw folder

- if the first image empty (not a test image taken after an abort): delete its corresponding .xml and run python /users/opd29/workspace/edna/bioSaxsv1/bin/ *.xml in raw folder

Note: better if not other EDNA process is running, particularly in HPLC case!


Known Issues

- Occasionally DAMMIN does not finish properly. This can cause error messages from EDNA1 (=slavia) and delay in HPLC processing. Solution: log in on slavia, use top to find the PID of dammin and kill -12 PID