EVOLUTION OF INFORMATION IN TRANSCRIPTIONAL REGULATORY SYSTEMS

Abstract

Transcriptional regulatory systems are a fundamental mechanism for regulating gene expression in all living organisms. Their implementation requires transcription factors (TFs) to bind specific target sites on the genome. In this thesis, the recognition of target sites by their cognate TFs is studied as a communication process, where the DNA sequences of the target sites encode information about TF-binding affinity decoded by the TF. Our approach, based on an ‘evolutionary information theory’ of biological codes, allows us to advance our understanding of how regulatory information is encoded and where it can evolve on DNA molecules. In this work, we reveal general principles that regulatory codes follow for the encoding of information as recognizable sequence patterns, or combinations of sequence patterns, and we present a novel algorithm to identify arbitrarily complex patterns in DNA sequences (Chapters 2-4). We also investigate the possibility that transcriptional regulation involves genetic elements that belong to distinct biological entities, forming ‘hybrid’ transcriptional regulatory networks whose information is encoded on independent DNA replicons. We provide evidence for the prevalence of this phenomenon, and we present a computational platform that can systematically uncover instances of such ‘hybrid’ gene networks (Chapters 5-6). Overall, this work contributes to the study of molecular evolution, showing that the genetic elements coding for transcriptional regulation follow specific information-encoding strategies that provide an advantage in terms of mutational robustness and evolvability. It also paves the way for a comprehensive study of gene networks that extends beyond the boundaries of a single genome.