Evolutionary Computation in Gene Regulatory Network Research

CHAPTER 4BICLUSTERING ANALYSIS OF GENE EXPRESSION DATA USING EVOLUTIONARY ALGORITHMS

Alan Wee-Chung Liew

School of Information and Communication Technology,Griffith University, Queensland, Australia

4.1 INTRODUCTION

The major goal of systems biology is to reveal how genes and their products interact to regulate cellular process. To achieve this goal, it is necessary to reconstruct gene regulatory networks, which help us to understand the working mechanisms of the cell. To infer the gene regulatory networks, one often looks at how groups of genes are co-expressed under certain conditions and how they regulate each other. This requires the use of high-throughput technologies such as whole genome expression profiling.

DNA microarray technologies allow us to have an insight into cellular process by simultaneously measuring expression levels of thousands of genes under various conditions. In a typical gene expression matrix, the rows describe genes and the columns describe conditions of the experiments. In DNA microarray experiments, discovering groups of genes that share similar transcriptional characteristics is instrumental in functional annotation, tissue classification, motif identification, and gene regulation [1, 2, 3]. Cluster analysis can help elucidate the regulation (or co-regulation) of individual genes, and therefore has been an important tool in gene regulation network study and network reconstruction [3]. However, in many situations, an interesting cellular process ...