2002 Volume 42 Issue 3 Pages 110-115
Advances in microarray techniques have enabled us to simultaneously monitor the expression levels of thousands or tens of thousands of genes under multiple conditions. Here, we present a novel method to analyze a large amount of gene expression profiles by the combination of the standard statistical techniques. The method is composed of two parts : the automatic determination of cluster boundaries in a hierarchical clustering and the following inference of a network by application of graphical Gaussian modeling. The feasibility of our method is discussed in terms of the biological knowledge of gene regulatory systems and the other analytical approaches.