A Large-scale Batch-learning Self-organizing Map for Function Prediction of Poorly-characterized Proteins Progressively Accumulating in Sequence Databases : Annual Report of the Earth Simulator Center April 2007 - March 2008

As a result of decoding of extensive genome sequences, a large number of proteins whose function cannot be predicted by the homology search of amino acid sequences is progressively accumulated and thus remains of no use in science and industry. A method to predict the protein function that does not depend on the sequence homology search is in urgent need. We previously developed a Batch-Learning SOM (BLSOM) for genome informatics, and in the present report, we describe use of the BLSOM method for prediction of protein function on the basis of similarity in composition of oligopeptides (di-. triand tetrapeptide in this study) of proteins. Oligopeptides are elementary components of a protein and involved in formation of functional motifs and structural organization of proteins. BLSOM for oligopeptides could extract characteristics of oligopeptide composition actualizing protein structure and function and thus predict functions.