Archived: rabbit proc

ARCHIVED. THIS IS OLD INFO FOR WHEN WE WERE USING PETER’S SCRIPTS (“RABBIT PROC”)

This is what happens automatically “in the background” while you are collecting data:

  • The beamline interface sofware, b3, communicates all snapshots and datasets it collects to a central database
  • An external data processing server monitors this database and finds out if there are any datasets collected that are in need of processing
  • If a ‘new’ dataset has been found that needs to be processed, a processing job is executed
  • Data processing is performed by XIA2 (http://xia2.blogspot.com; Acta Cryst D 2013 Jul 1; 69(Pt 7): 1260–1273)
  • In order to make it more robust, data is processed as follows:

A. Try xia2 with XDS and labelit indexing on 2 frames

B. If A fails, try xia2 with XDS using indexing with all frames

C. If B fails, try processing with DIALS

D. If C fails, try with MOSFLM

E. If D fails, give up

  • Once the data is processed, it is moved into the users home directory under ~/rabbit_proc

WHAT TO EXPECT:

  • The processing will try and make sure that your data is named in a unique way. The way it does this is by placing the results in the following directory structure: /your_home_directory/rabbit_proc/data_set_name_beamline_year_date_time
  • Within the directory, you will find the xia2 output structure and some additonal files. In particular, the scaled reflections file is in the subdirectory DataFiles/P_X_scaled.sca

WHAT TO KEEP IN MIND

  • The data processing is an automated process. It doesn’t hurt to check if the results are reasonable. In the vast majority of cases it will be.
  • Don’t call your dataset something very generic, use logical names that are project specific, it will make browsing your results later on a lot easier.
  • Data processing can take anywhere from 10 minutes to 2 hours, depending on the quality. Typically you should expect to see something appearing within 20 minutes. (We are working on speeding this up.)