Gene Regulatory Networks

Abstract

Two gene regulatory networks inferred from different types of data are considered in this chapter. Gene expression networks are networks inferred from microarray time series data and transcription factor networks are networks obtained from a new genome-wide technique that allows an identification of all of the DNA binding sites for each transcription factor (TF). While addressing the same underlying questions, these networks reflect different properties of gene regulation and provide different insights. The gene expression network is inferred from dynamic analysis of time series data of gene expression profiles. The TF net-works, on the other hand, are a direct result of experimental observation of a physical association between a TF and a DNA binding site, which (except for experimental noise) is unique. While our knowledge of the transcription factor networks is limited, these networks provide insights into a regulatory core network of TFs that regulate each other, and drive all network interconn ectivity. In both cases, the resulting networks show features that may be universal to biological systems. The global properties of such networks show the scale-free distributions of node connectivity indicative of a hierarchical network and also exhibit small world graph properties. We discuss a network growth model based on gene duplication that provides excellent agreement with the global network parameters derived from the analysis of experimental expression data. In addition to these global properties, the local properties of these gene expression networks can be used in data mining and classification.