GenExplore: interactive exploration of gene interactions from microarray data

Abstract

DNA microarray provides a powerful basis for analysis of gene expression. Data mining methods such as clustering have been widely applied to microarray data to link genes that show similar expression patterns. However, this approach usually fails to unveil gene-gene interactions in the same cluster. We propose to combine graphical model based interaction analysis with other data mining techniques (e.g., association rule, hierarchical clustering) for this purpose. For interaction analysis, we propose the use of graphical Gaussian model to discover pairwise gene interactions and loglinear model to discover multigene interactions. We have constructed a prototype system that permits rapid interactive exploration of gene relationships.