Prediction of protein-protein interaction based on structure.

A great challenge in the proteomics and structural genomics era is to predict protein structure and function from sequence, including the identification of biological partners. The development of a procedure to construct position-specific scoring matrices for the prediction and identification of sequences with putative significant affinity faces this challenge. The local and web applications used for sequence and structure search, sequence alignment, protein modeling, molecule edition and modification, and scoring matrices construction are described in detail. The methodology is based on the information contained in structural databases and takes into account the subtle conformational and sequence details that characterize different structures within a family. Using the matrices, the protein sequence databases can be easily scanned to locate putative partners of biological significance. The success of this methodology opens the way for the prediction of protein-protein interaction at genome scale.