[phenixbb] Raw diffraction images for SARS-CoV-2 related structures

Gerard Bricogne gb10 at globalphasing.com
Wed Mar 18 15:33:59 PDT 2020

Dear colleagues,

Perusal and some initial (re-)refinement of the various SARS-CoV-2 protease
structures in the PDB seems to indicate that that there might be potential
to improve these if refinements could be repeated after some reprocessing
and further analysis of the raw diffraction images, rather than against the
deposited merged data. This statement should in no way be construed as a
criticism of the remarkable achievements of the research groups concerned,
who have been operating under tremendous time pressure, but as an exciting
opportunity to push methods to their limits on a uniquely significant class
of structures.

Another consideration is that the various logistical problems created by
COVID-19 may soon make it increasingly difficult to collect new diffraction
data on potential drug targets relevant to the fight against SARS-CoV-2,
underlining the importance of ensuring that the best results be obtained
from every dataset actually collected, and that the most useful conclusions
be drawn from the analysis of those datasets towards improving the quality
of subsequent data collections. 

On this basis we would like to propose that special efforts be made to grant
public access to the raw image data associated with any SARS-CoV-2 related
structure that is deposited into the PDB. This can be done by (1) archiving
these raw image data using resources such as data.sbgrid.org, zenodo.org,
proteindiffraction.org or any other cloud-based data-sharing service, and
(2) communicating the corresponding DOIs to the wwPDB centres. This idea
could be extended to datasets that investigators would like to offer to
interested methods developers or expert users at the pre-deposition stage.

Experts making use of those raw data would be encouraged to document, in as
much detail as possible, how particular programs or workflows could be used
on those structures/datasets to obtain the best results. This would be a
kind of "virtual workshop", a particularly valuable collective activity at
the present time when several in-person workshops (e.g. RapiData) have been
cancelled and many meetings are in limbo for several months.

The latter activity would benefit from having a centralised facility set up
for the experts to post their results and annotations: we could create such
a facility, but other, larger groups might want to consider doing so. 

With best wishes,

Clemens & Gerard.

More information about the phenixbb mailing list