结构生物学sturcture biology explore-68.pdfVIP

28 Prediction of protein–protein interactions by docking methods Graham R Smith* and Michael JE Sternberg*† Recently, developments have been made in predicting the been, and continues to be, targeted towards those complexes structure of docked complexes when the coordinates of the for which the conformational change on docking is fairly components are known. The process generally consists of a small, thus enabling the use of methods that are based on stage during which the components are combined rigidly and shape and chemical complementarity of the unbound then a refinement stage. Several rapid new algorithms have components. This can be considered the ‘canonical case’ of been introduced in the rigid docking problem and promising protein–protein docking and will be the focus of this review. refinement techniques have been developed, based on modified molecular mechanics force fields and empirical The character of the interfaces in known protein–protein measures of desolvation, combined with minimisations that complexes has been thoroughly investigated in several switch on the short-range interactions gradually. There has also reviews [1–3]. Whilst obligate complexes do have an been progress in developing a benchmark set of targets for interface that is slightly more hydrophobic than the rest docking and a blind trial, similar to the trials of protein structure of the molecular surface, there is very little indication of prediction, has taken place. this in nonobligate complexes. The subset of complexes for which both components have been crystallised separately Addresses unbound provide