Data-driven inference of dynamic transcriptional regulatory mechanisms in prokaryotes: a systems perspective

Abstract

Microbes tailor their physiology to diverse environments despite having streamlined genomes and few regulators. Mechanisms by which microbes expand their genetic repertoire include modular reorganization of genetic expression through dynamic activity of complex gene regulatory networks (GRNs). Deciphering accurate GRNs is essential to understand how their topology contributes to cellular behavior. This dissertation develops computational methods to reverse engineer GRNs directly from genome sequence and transcriptome data. These data-driven models capture dynamic interplay of environment and genome-encoded regulatory programs for two phylogenetically distant prokaryotes: <italic>E. coli</italic> (a bacterium) and <italic>H. salinarum</italic> (an archaeon). The models reveal how distribution of cis-acting gene regulatory elements (GREs) and condition-specific influence of transcription factors (TFs) at each element produces environment-specific transcriptional responses. These regulatory programs partition and re-organize transcriptional regulation of genes within regulons and operons into condition-specific co-regulated modules, or corems. Corems capture fitness-relevant co-regulation by different transcriptional control mechanisms acting across the entire genome. Organization of genes in corems defines a system-level principle for prokaryotic gene regulatory networks that extends existing paradigms of gene regulation and helps explain how microbes negotiate environmental change.