Algorithms in Computational Molecular Biology: Techniques, Approaches and Applications

CHAPTER 41

UNSUPERVISED LEARNING FOR GENE REGULATION NETWORK INFERENCE FROM EXPRESSION DATA: A REVIEW

Mohamed Elati and Céline Rouveirol

41.1 INTRODUCTION

In the organism, each organ, each cell, and each protein has a defined role to fulfill so that life is maintained. When one player deviates from its predefined scenario, the consequences are generally not even noticeable. However, sometimes, the consequences are very important, and the whole organism might not survive it. The only way to fight against these disorders is to understand fully the functioning of the whole ensemble, and how each deviation affects it. This is what systems biology is about [2, 14]. At the cellular level, this requires understanding how the different components interact, and how these interactions result in functional networks, often called pathways. Regulation of gene expression at the transcriptional level is a fundamental mechanism that is evolutionarily conserved in all the cellular systems. This form of regulation is typically mediated by transcription factors (TFs) that bind to short DNA sequence motifs, also called binding sites, in promoter regions of transcriptional units and either activate or repress the expression of nearby genes.

In the past decade, microarrays and other high-throughput methods have revolutionized biology, providing a wealth of new experimental data. This unprecedented amount of data—for example, the simultaneous measurement of the expression levels of all genes in a given ...