In this paper, we evaluate a method for analyzing microarray data. The method is an attempt to learn regulatory interactions between genes from gene expression data. It is based on a Bayesian network, which is a mathematical tool for modeling conditional independences between stochastic variables. We review the dynamic nature of interacting genes, and explain how to model them using such a network. We also discuss an algorithm for learning the structure of interacting genes from gene expression data, using scaled mutual information for scoring.