UC San Diego

Systems analysis of cell-scale reconstructions has become a staple of modern biological discovery in the era of high -throughput data generation. Reconstructions of bacteria currently assimilate the existing wealth of biochemical knowledge for a given organism, accounting for metabolites and the stoichiometry of their transport and transformation via enzymes. However, the bulk of the material and energy consumed during cellular growth is devoted to the synthesis of the macromolecules involved in information transfer: RNA, protein, and the genome itself. In this dissertation, a framework for incorporating the fundamental processes of bacterial transcription and translation is described. This framework is used to integrate mRNA expression and half-life data towards assessing the global transcription state of the Escherichia coli genome under numerous experimental conditions. To address how the translation network would be constrained within this framework, a model is used to determine the theoretical limits in translational efficiency achievable by altering the synonymous codon usage of each gene given measured tRNA abundances. A sensitivity analysis of translational efficiency, which can be varied by an average 6.5-fold, demonstrates that wild-type synonymous codon usage and measured tRNA abundances in E. coli are highly synchronized. However, the results from these studies also expose the limitations of network reconstructions that neglect three-dimensional spatial information. The transcription state of a genome can be highly nonrandom with respect to chromosomal position. Furthermore, tRNA diffusion limitations must be taken into account to accurately model translational efficiency. This dissertation thus advances a method for characterizing the spatial organization of chromosome position-dependent data. A comprehensive assessment of the periodic pattern content contained within 163 prokaryotic chromosomes is performed using wavelet analysis. The degree of patterning in sequence-derived properties correlates with genome-size, overall GC-content, and the occurrence of motility and chromosomal-binding proteins. Given additional functional data for E. coli, long-range patterns in multiple heterogeneous properties are shown to be highly correlated and are consistent with experimentally detected chromosomal macrodomains. Taken together, the findings reported in this dissertation demonstrate that the field of cell-scale modeling will ultimately enter a phase in which network connectivity is viewed within the context of topobiological, spatial constraints