结构生物学sturcture biology explore-33.pdfVIP

368 Computational methods for the prediction of protein interactions Alfonso Valencia* and Florencio Pazos† systems, such as yeast [2• •• •• • Establishing protein interaction networks is crucial for ,3,6 ,7 ] and Helicobacter pylori [8 ]. understanding cellular operations. Detailed knowledge of the And, although the limits of resolution of these approaches ‘interactome’, the full network of protein–protein interactions, are open to discussion [9,10••], they do nevertheless promise in model cellular systems should provide new insights into the much for the future. structure and properties of these systems. Parallel to the first massive application of experimental techniques to the In parallel, a number of computational methods have been determination of protein interaction networks and protein developed for the prediction of protein interactions from complexes, the first computational methods, based on genomic information [11,12•], extending into the prediction sequence and genomic information, have emerged. of the residues that participate in the interacting surfaces. Here, we describe the five computational techniques Addresses available for the prediction of interaction partners and *Protein Design Group, National Center for Biotechnology, examine their range of applicability. In addition, we analyze CNB-CSIC, Cantoblanco, 28049 Madrid, Spain; new trends in the determination of interacting surfaces on e-mail: valencia@cnb.uam.es †ALMA Bioinformática, Centro Empresarial Euronova, Ronda de Poniente, the basis of s