Mining Gene Expression Profiles with Biological Prior Knowledge

One of the important goals in the post-genomic era is to identify the functions of genes, either individually or as group. Recently, there has been an increasing use of the gene ontology (GO) to analyze a list of genes identified via various statistical and/or computational methods. The main assumption behind using GO for interpreting microarray data is that the genes that belong to similar molecular functions or biological processes would display similarly tightly regulated expression patterns. Current methods utilize GO after the statistical analysis of gene expression data. In this paper, we describe a method that utilizes both gene expression values and biological knowledge simultaneously to identify the significant biological functions. The method is different from other methods in that it incorporates GO as prior knowledge into the mining of gene expression data. The method has been applied to the gene expression profiles to cell cycle experiments