Semisynthetic Multi-Lattice Diffraction Data

Semisynthetic Multi-Lattice Diffraction Data                

Richard Gildea, Graeme Winter*, Diamond Light Source

*graeme.winter@diamond.ac.uk

DOI: 10.5281/zenodo.10820

These data are published in support of "New methods for indexing multi-lattice diffraction data" manuscript submitted to Acta Crystallographica section D, authors:     

Richard J. Gildea (1), David G. Waterman (2, 3), James M. Parkhurst (1), Danny Axford (1), Geoff Sutton (4), David I. Stuart (1,4), Nicholas K. Sauter (5), Gwyndaf Evans (1) and Graeme Winter (1)

(1) Diamond Light Source (2) STFC Rutherford Appleton Laboratory (3) Research Complex at Harwell (4) Division of Structural Biology, Wellcome Trust Centre for Human Genetics (5) Lawrence Berkeley National Laboratory

for the community of methods developers and interested persons to test other algorithms.

This work is licensed by Diamond Light Source Ltd under the Creative Commons Attribution 4.0 International Licence (CC-BY):                              

https://creativecommons.org/licenses/by/4.0/

Creation of the Data
Bovine pancreatic trypsin crystals were prepared following standard methods (thanks to Carina Lobley @ Diamond Light Source for these) and small wedges of data were taken on beamline I04 at Diamond Light Source during in-house time (thanks to David Hall @ Diamond Light Source for this).               

Each original data set a - l was collected with arbirtary kappa angles, a total of 10 degrees of rotation with 0.1 degree increments. These data were then summed pixel-wise as follows:

To make two-lattice set ag, the counts on every pixel i, j on every image k were added from sweep a and sweep g to get the pixel count for semisynthetic sweep ag. Thus the sweeps a - l were combined to give 12 x 1 lattice, 6 x 2 lattice, 4 x 3 lattice, 3 x 4 lattice and 2 x 6 lattice example sets, which may be found as follows:                                                     

1:
a  b  c  d  e  f  g  h  i  j  k  l

2:
ag  bh  ci  dj  ek  fl

3:
aei  bfj  cgk  dhl

4:
adgj  behk  cfil

6:
acegik  bdfhjl

Clearly, as the pixels are added more peaks will appear but the background will also increase. The structure of these data allows the original images to be processed as a reference for the initial data quality and the combinations to be processed to assess the effect on the quality of superimposing the data.

Methods

The data were added as follows:

• each image was read, the pixel data to a flex array and the header to a string
• the pixel data in the flex arrays added to make the n-lattice image
• this array re-compressed using the CBF byte-offset compression, added to the existing header string and written to disk
• this procedure was followed for all images in a sweep

Python code for these procedures is available on request from the authors.