Evolutionary Computation in Gene Regulatory Network Research

CHAPTER 12ARTIFICIAL GENE REGULATORY NETWORKS FOR AGENT CONTROL

Sylvain Cussat-Blanc, Jean Disset, Stéphane Sanchez and Yves Duthen

University of Toulouse – IRIT – CNRS UMR5505, Toulouse, France

12.1 INTRODUCTION

Gene regulatory networks (GRNs) are biological structures that control the internal behavior of living cells. They regulate gene expression by enhancing and inhibiting the transcription of certain parts of the DNA. However, they can be used as agent controllers: for example, instead of regulating gene expressions, they can be used to regulate either agent actuators or high-level complex behaviors. This chapter summarizes three applications of a computational model of gene regulatory network to control virtual agents. The aim is to provide an overview of problems that gene regulatory networks can address in reference to some recent works done in this area.

When used to simulate gene expression regulation, a GRN is usually encoded within a bit string, as DNA is encoded within a nucleotide string. As in real DNA, a gene sequence starts with a particular sequence, called the promoter in biology [19]. In the real DNA, this sequence is represented with a set of four proteins: TATA where T represents the thymine and A the Adenine. Torsten Reil is one of the first to propose a biologically plausible model of gene regulatory networks [21]. The model is based on a sequence of bits in which the promoter is composed of four bits 0101. The gene is coded directly after this promoter ...