traditional, glass-fibre, mounts to the use of magnetic bases fitted with Kapton
micro-mounts. Crystals are usually bonded to the micro-mounts with a suitable
inert oil and handled in a manner that depends on the specific properties of the
sample, whether that is due to a possible chemical reaction or loss of solvent on
exposure to the air. By immediately flash-freezing the crystals under liquid nitrogen
in a Unipuck container, the samples can then be couriered to the synchrotron,
ice-free, with the Unipucks held within the inert conditions of a dry shipping
dewar. Each Unipuck can contain 16 sample mounts, and, with each dry dewar
able to hold 7 Unipucks, a single shipment of 112 crystals can be made. On receipt of
the dry shipper at the beamline, all 7 Unipucks can be loaded into the dewar of the
robotic sample changer. Johnson et al. have found that distributing a number of
crystals of the same sample in different pucks increases the chance of selecting a
sample suitable for structure determination. This approach reduces the risk of losing
the opportunity to collect data on a particular sample, in the event of an issue with an
individual puck, as the spread of crystals between alternative pucks provides some
redundancy.
Data Collection Remote data collection is carried out by the users connecting to the
beamline using the NoMachine remote desktop software, and this can be carried out
from the users’ home laboratory (or indeed, with suitably fast broadband, from their
home). The arrangement provides a very similar experience to the user being present
in the control room of the beamline, although it is recommended that a workstation
with at least two large screens is used due to the need to display several large
software windows for General Data Acquisition (GDA), Information System for
Protein Crystallography Beamlines (ISPyB) and some Experimental Physics and
Industrial Control Systems (EPICS) controls. A key difference to the operation of the
beamline GDA is the use of a software ‘Baton’, designating which computer
currently holds control of the diffractometer and robot. This must be passed explicitly from the beamline to the user, with the local beamline staff able to take back
control, via the baton, at any point. To aid communication between beamline staff
and the remote users, a text messenger is incorporated within the GDA.
Data Processing With the more efficient use of the beamline and the corresponding
increased rate at which data are recorded, the degree of data analysis that can be
carried out during a session of beamtime can only be relatively rudimentary as there
is time pressure to collect data from other crystals. The automated data processing
pipeline can greatly assist in providing reduced data, and solved crystal structures,
for well-behaved samples, but for more challenging cases, data analysis will require
additional effort. For these more difficult datasets, which may involve twinning
(or where there are multiple-crystal components) and major structural disorder, the
diffraction images can be either imported directly into or converted into the appropriate file formats alternative processing packages that fit with the users’ previous
experience and preference. These software packages, whether they are public
domain or proprietary, contain more bespoke tools for small molecule crystallography that allow useful integrated data to be produced from the usually very poorly
diffracting crystals that are brought to the synchrotron.
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
87
micro-mounts. Crystals are usually bonded to the micro-mounts with a suitable
inert oil and handled in a manner that depends on the specific properties of the
sample, whether that is due to a possible chemical reaction or loss of solvent on
exposure to the air. By immediately flash-freezing the crystals under liquid nitrogen
in a Unipuck container, the samples can then be couriered to the synchrotron,
ice-free, with the Unipucks held within the inert conditions of a dry shipping
dewar. Each Unipuck can contain 16 sample mounts, and, with each dry dewar
able to hold 7 Unipucks, a single shipment of 112 crystals can be made. On receipt of
the dry shipper at the beamline, all 7 Unipucks can be loaded into the dewar of the
robotic sample changer. Johnson et al. have found that distributing a number of
crystals of the same sample in different pucks increases the chance of selecting a
sample suitable for structure determination. This approach reduces the risk of losing
the opportunity to collect data on a particular sample, in the event of an issue with an
individual puck, as the spread of crystals between alternative pucks provides some
redundancy.
Data Collection Remote data collection is carried out by the users connecting to the
beamline using the NoMachine remote desktop software, and this can be carried out
from the users’ home laboratory (or indeed, with suitably fast broadband, from their
home). The arrangement provides a very similar experience to the user being present
in the control room of the beamline, although it is recommended that a workstation
with at least two large screens is used due to the need to display several large
software windows for General Data Acquisition (GDA), Information System for
Protein Crystallography Beamlines (ISPyB) and some Experimental Physics and
Industrial Control Systems (EPICS) controls. A key difference to the operation of the
beamline GDA is the use of a software ‘Baton’, designating which computer
currently holds control of the diffractometer and robot. This must be passed explicitly from the beamline to the user, with the local beamline staff able to take back
control, via the baton, at any point. To aid communication between beamline staff
and the remote users, a text messenger is incorporated within the GDA.
Data Processing With the more efficient use of the beamline and the corresponding
increased rate at which data are recorded, the degree of data analysis that can be
carried out during a session of beamtime can only be relatively rudimentary as there
is time pressure to collect data from other crystals. The automated data processing
pipeline can greatly assist in providing reduced data, and solved crystal structures,
for well-behaved samples, but for more challenging cases, data analysis will require
additional effort. For these more difficult datasets, which may involve twinning
(or where there are multiple-crystal components) and major structural disorder, the
diffraction images can be either imported directly into or converted into the appropriate file formats alternative processing packages that fit with the users’ previous
experience and preference. These software packages, whether they are public
domain or proprietary, contain more bespoke tools for small molecule crystallography that allow useful integrated data to be produced from the usually very poorly
diffracting crystals that are brought to the synchrotron.
Leading Edge Chemical Crystallography Service Provision and Its Impact on. . .
87