CCP4: Low resolution complexes; handling difficult data; empowering structural biologists and supporting UK structural biology

Structural Biology is a powerful tool for analysing biological molecules. In essence it locates the atoms that make up proteins and other biomolecules. It is used in basic science to define the 3D structures of such molecules as the ribosome, ion channels and complex protein assemblies such as ATPase. Once a 3D 'model' of all the atoms in the biological model is available, it is possible to make deductions about how the molecule functions within the organism. This knowledge has corresponding applications in controlling that function, often leading to medical advances in drug design, better understanding of biological and feedback systems in the natural environment, new approaches in chemical engineering, and many other benefits. The process of 'solving' the structure involves first obtaining crystals and then exposing these to X-rays: this is equivalent to using a very powerful microscope and allows us to 'see' individual atoms. Getting crystals is in itself very demanding and involves a lot of skill and scientific insight. There is an additional problem in using X-rays in that there is no lens system for the microscope. This means that indirect, so-called diffraction images of the crystal are recorded. The process of extracting data from these images and proceeding to a final structure is computationally intensive and requires an entirely different set of skills from those of the biologist. Developing new computer methods and crucially embedding them in robust easy-to-use software will transform structural biology from a labour intensive highly technical process to a routine tool in biology. The main focus of this grant is on the computational process of solving 3D structures. CCP4 is a project which has collected, developed, packaged and distributed software for the many stages of this process, acting as a focus for methods development for over 25 years and making software available to both Academic and Industrial teams working in this area. It is an excellent example of technology transfer of basic science into a major UK Industry: Pharmaceutical development. The problem cannot yet be said to be totally solved. Biologists firstly are tackling more and more difficult structures, which require novel techniques. In addition, they are solving large number of sometimes related structures, such as a series of mutants and complexes, where automated high throughput is vital and systematic record keeping is essential. This proposal is focussed on enabling these developments.

1. RA Post 1. Low resolution structures present special problems. Template-matching methods developed by Dr Cowtan provide a proven basis for the interpretation of maps of limited quality, and these methods will be enhanced by the use of larger scale structural information; this approach will be applied to model building and refinement at lower resolutions. 2. RA Post 2. Integrating intensities from diffraction images represents a crucial and fundamental step in the structure determination process, as all subsequent steps are dependent on these measurements and their estimated standard uncertainties. New data processing algorithms will be developed to handle multiple lattices, to automate processing and data analysis and to deal with the new Pilatus detectors. 3. RA Post 3. A database back-end will be developed to systematise and store all the data relevant to structure solution. We will design sensible 'wizards' to improve the consistency and accelerate the rate of structure determination, based on systematic handling of this information. 4. RA Post 4. Finally, for these tools to have the maximum benefit they will be translated into robust well-supported software distributed to the community. Software produced by tasks 1-3, by other CCP4 supported projects, and by independent contributors will be integrated into the suite in a way that is intuitive for a non-specialist. These major tasks will be complemented by suite-wide changes that keep CCP4 relevant to the latest challenges, such as large complexes.