Gene clustering and gene function prediction using multiple sources of data

Gene function prediction and gene clustering using biological information, including genome sequence, gene expression data, protein interaction data, phylogenetic data, etc., is an important step toward the inference of the gene regulatory network in the cell. Different types of data reveal different aspects of the relationships among the genes within a set. It is expected that each type of data has its own strengths and weaknesses in discovering specific relationships. We propose a new method to optimally cluster genes and to predict the function of unknown genes based on multiple sources of data by maximizing the total similarity gain function within all clusters.