Work

Data collection and processing

Year: 2013
Type: article
Abstract: In 2012 the title for the annual CCP4 Study Weekend was Data Collection and Processing, following the tradition of revisiting major subjects in macromolecular crystallography.The conference was held at the University of Warwick and attended by more than 400 scientists, mainly from the UK but also a significant number from further afield.Data collection is the last experiment in an often arduous pursuit of determining the three-dimensional structure of a macromolecule.A very large amount of time, effort and hard-to-get funding has generally been spent on obtaining sufficient quantities of purified material, and crystals of sufficient quality.Exposing the crystal to X-rays, either using a home source or a synchrotron, and obtaining the best possible data should therefore not be taken lightly.Surprisingly, this final experiment is often carried out hastily and without due attention, simply taking an approach that has become habitual.The disappointment felt when a data set turns out to be incomplete, or of insufficient quality to solve or refine the structure, is enormous and is an excellent reason to study the proceedings in this issue with the attention they deserve.In recent years there have been several major developments in data collection and processing.Third-generation synchrotron sources are now widely available, sample handling has been automated at most facilities, and pixel array detectors have made it possible to perform shutterless data collections, reducing errors and allowing data sets to be collected in mere minutes.At the same time data-processing software has been streamlined, and expert software packages have become available that in many cases are able to analyse the raw images and provide the user with data reduced in the correct space group.It is paramount that these developments are used properly, so as to optimize the use of large scale facilities such as synchrotrons and more recently free electron lasers.At the same time it gives the user more time to carefully consider how to utilize all these tools in order to obtain the best data.The meeting was opened with introductory talks by Zbyszek Dauter and James Holton.Zbyszek covered the basics of data collection, including the consideration of symmetry, unit-cell alignment in the beam, crystal lifetime, how to avoid overlaps and overloads and many other fundamental concepts that need to become second nature.He stressed that how to collect data depends on what the data are needed for, e.g.native data or data for molecular replacement or phasing calculations.James re-iterated that the ratio of data sets collected to structures deposited remains surprisingly large and went on to warn about overlaps and radiation damage.The last part of his talk concentrated on 'signal to noise', and what can be done to optimize this.After lunch there was a series of talks about data processing and data quality.Andrew Leslie covered the theory and practice of autoindexing, explaining the underlying principles and highlighting what parameter choices are important for a successful outcome.Jim Pflugrath described the program d*TREK.Phil Evans and Kay Diederichs then discussed aspects of data quality.Both highlighted the use of CC 1/2 as a far better measure of data quality, and a criterion for determining the resolution limit of the data.The acceptance of such better measures of data quality than the R merge is not yet universal, but several recent publications highlighting their usefulness should improve this situation.The tea break was followed by four talks on more technical aspects of data collection.Jan-Pieter Abrahams spoke about progress in the collection and processing of electron diffraction data.Thomas White described the sample handling when using the free electron laser as a radiation source where a suspension of crystals is injected into the beam path, ensuring a continuous supply of fresh crystals.A very large number of data images are collected, and the approach taken to processing these was described in some detail.Andrew McCarthy detailed the use of a mini-kappa goniometer head and the associated software as implemented at the ESRF.Using several examples he showed how
Language: English
Open Access status: hybrid
Estimated APC paid: $4,740
OpenAlex ID: W2044099130
FWCI: 0.131
Cites: 4
Cited by: 153
Related to: 10