Bayesian probabilistic approach for predicting backbone structures in terms of protein blocks

By using an unsupervised cluster analyzer, we have identified a local struc tural alphabet composed of 16 folding patterns of five consecutive C-alpha ("protein blocks"). The dependence that exists between successive blocks is explicitly taken into account. A Bayesian approach based on the relation p rotein block-amino acid propensity is used for prediction and leads to a su ccess rate close to 35%, Sharing sequence windows associated with certain b locks into "sequence families" improves the prediction accuracy by 6%, This prediction accuracy exceeds 75% when keeping the first four predicted prot ein blocks at each site of the protein. In addition, two different strategi es are proposed: the first one defines the number of protein blocks in each site needed for respecting a user-fixed prediction accuracy, and alternati vely, the second one defines the different protein sites to be predicted wi th a user-fixed number of blocks and a chosen accuracy, This last strategy applied to the ubiquitin conjugating enzyme (alpha/beta protein) shows that 91% of the sites may be predicted with a prediction accuracy larger than 7 7% considering only three blocks per site. The prediction strategies propos ed improve our knowledge about sequence-structure dependence and should be very useful in ab initio protein modelling. (C) 2000 Wiley-Liss, Inc.