Computational analysis and prediction of protein-protein interactions

Protein-protein interactions perform and regulate fundamental cellular processes. The comprehensive study of such interactions on a genome-wide scale will lead to a clearer understanding of diverse cellular processes and of the molecular mechanisms of disease but interactions are difficult to determine reliably by experimental methods alone. Over the last five years we have developed effective computational methods to predict protein-protein interactions in human by combining data from orthology, expression, posttranslational modification, domains and gene ontology terms within a Bayesian framework (1,2). This project will build on this solid foundation to extend the method to more species and to explore alternative training methods to Bayes. For example, combination of features by Artificial Neural Networks or Support Vector Machines. Predictions will be made for protein familes of interest to colleagues in the College of Life Sciences and complementary experimental data sought for strong positive predictions. The student on this project will have ample opportunity to develop strong skills in database programming, statistics, software development and data analysis. The project will also investigate the use protein-protein interaction networks to infer functional clusters that are not obvious from simple binary interaction predictions.