The talk will present recent work on protein modelling covering the following topics: i) a novel approach for ab initio protein structure prediction; ii) analysis and prediction of protein function and iii) studies on protein docking.
The prediction of a protein structure from sequence in the absence of a homologue of known conformation to act as a template remains an unsolved problem. Mr Ben Jeffreys and Dr Lawrence Kelley in my group have recently developed a rapid novel ab initio folding approach. A protein is represented as a single side-chain sphere attached to a C alpha position. Springs are used to model thermodynamic effects and Langevin dynamics is used to fold the protein from an open chain to a tertiary structure. Current results are yielding international quality predictions in hours of CPU time.
Up to 50% of the proteins in a genome lack functional annotation. We report a novel approach [1] developed to predict protein function from sequence (ConFunc) which uses the gene ontology annotation to supervise the construction of profiles specific for protein function. ConFunc is available to the community via a web server. ConFunc and many other function prediction methods rely on homology for function. A recent analysis [2] describing the relative roles of convergent and divergent evolution of enzyme function will be reported.
A new approach, 3D-Garden [3] to predicting the structure of a protein complex starting from two unbound components will be described. The approach is based on matching the two surfaces described as marching cubes. Scoring is via an all atom potential and the approach is being developed to include side-chain and main-chain flexibility. In predictive docking it would be most helpful to know if the protein were flexible docking needs to be included in the modelling. A recent study [4] using a coarse grained elastic normal mode simulation has shown that the frequency spectrum of normal modes can be predict if a protein will not undergo substantial conformational change on docking.
[1] Wass, M.N. and Sternberg, M.J. (2008) ConFunc - functional annotation in the twilight zone, Bioinformatics, 24, 798-806.
[2] Gherardini, P.F., Wass, M.N., Helmer-Citterich, M. and Sternberg, M.J.E. (2007) Convergent Evolution of Enzyme Active Sites Is not a Rare Phenomenon, J Mol Biol, 372, 817-845.
[3] Lesk, V.I. and Sternberg, M.J. (2008) 3D-Garden: a system for modelling protein-protein complexes based on conformational refinement of ensembles generated with the marching cubes algorithm}, Bioinformatics, 24, 1137-1144.
[4] Dobbins, S.E., Lesk, V.I. and Sternberg, M.J.E. (2008) Insights into protein flexibility: the relationship between normal modes and conformational change on protein-protein docking, Proc Natl Acad Sci U S A, in the press.