Abstract:
Proteins carry out most cellular processes as protein modules. Systematic identification of these protein functional modules provides essential knowledge linking proteome dynamics to cellular function and phenotype. This is one of the most challenging tasks at present since most genomes are successfully sequenced and genes identified.

We give a brief introduction to the rapid growing field of genomics and clarify the vital role of protein interaction studies. We then describe two graph algorithms for detecting protein modules: the spectral clustering, and clique/biclique finding. The spectral clustering formulates the problem as eigenvectors of the graph Laplacian; The maximal clique/biclique finding algorithms use a new type of constrained quadratic optimizations. Both algorithms have complexity of O(||E||) (# of edges) and can be efficiently computed on parallel architectures. The biological significance of the discovered protein modules are verified through concepts from the Gene Ontology. A number of uncharacterized proteins are found to be new members of important protein complexes.

Bio:
Chris Ding is a staff computer scientist at Lawrence Berkeley National Laboratory. He received a Ph.D. from Columbia Univerisity and did research at California Institute of Technology and Jet Propulsion Laboratory. His research focuses on bioinformatics and machine learning / data mining. He develops efficient graph algorithms using matrix computation.

More information about him can be found at http://crd.lbl.gov/~cding.