UC San Diego

In this dissertation, I develop ME-Models. ME-Models are genome-scale models that seamlessly integrate metabolic and gene product expression pathways. They can be used to compute optimal cellular states for growth in steady-state environments. They take as input the availability of nutrients to the cell and produce experimentally testable predictions for : (1) the cell's maximum growth rate [mu]* in the specified environment, (2) substrate uptake/by- product secretion rates at [mu]*, (3) metabolic fluxes at [mu]*, and (4) gene product expression levels at [mu]*. Unlike previous genome-scale models, ME-Models explicitly account for the production of all RNAs and proteins. I first build a prototype ME-Model for the simple microorganism Thermotoga maritima. The T. maritima genome was sequenced in 1999, and needed correction and complete re-annotation. I developed a framework drawing on multi- omic data to annotate genomic features involved in transcription, translation, and regulation. These features in T. maritima were found to display distinctive properties. In addition to basic characterization, the re- annotation was used to build the T. maritima ME-Model. Reactions to produce all the RNAs and proteins were added to its metabolic model, and metabolism was linked to gene expression through 'coupling constraints.' In the second part of this dissertation, the method was extended to E. coli. Backed by the wealth of phenotypic information available for this organism, I was able to firmly support the statement that ME-Models extend and refine microbial growth phenotype prediction. Next, a previous model predicted a ppc knockout of Salmonella enterica serovar Typhimurium would grow, but it did not experimentally. Ultimately, network modeling pinpointed the cause of the discrepancy (the inability of cells to route flux through the glyoxylate shunt when ppc is removed). The ppc project illustrates the importance of considering expression and regulation in genome-scale models. Finally, I demonstrate (albeit preliminarily) that ME-Models begin to bridge systems and synthetic biology approaches for engineering life