Incorporating biological knowledge in gene expression analysis


Rudy Guerra
Rice University

Abstract

            Gene expression data is typically analyzed for differential expression and clustering. The latter is in part performed in the hope of learning about the biological function of genes. The logical basis is that if two genes exhibit similar expression profiles (especially time course), then they share a common biological function.  Interestingly, clustered expression profiles provide a basis for defining multivariate response data that may improve power in searching for expression trait loci (ETL). Both clustering and the ETL problem can be aided by incorporating biological information for the genes under analysis.

        This talk will present on-going research in finding optimal procedures for incorporating the biological information, which in our applications come from gene ontology structures.