Independent component analysis and scoring function based on protein interactions

We describe an approach for discovering biological gene clusters from gene expression data of DNA microarray and scoring the genes based on protein interaction data. Our approach is based on the assumption that many clusters exhibit two properties, i.e., their genes exhibit a similar gene expression profile and the protein products of the genes often interact. Our approach to clustering is based on the independent component analysis model, which uses the ICA algorithm and our approach to scoring is based on number of protein product interactions of the genes within a cluster. We present the results on Saccharomyces cerevisiae gene expression dataset combined with a binary protein interaction data set.