Expanding the Scope of Genome-scale Models of Metabolism and Gene Expression
- Author(s): Lloyd, Colton Joseph
- Advisor(s): Palsson, Bernhard O
- et al.
The cost of whole genome sequencing has declined precipitously over the past two decades. This reduced cost of data collection has contributed to a deluge of multi-omics data types for Escherichia coli and other commonly studied microbes. As a result, methods to obtain actionable information from whole genome sequencing and other omics data have become increasingly valuable. Here, we broaden the scope of genome-scale models, thus allowing them to add context to multi-omics data and add insight into E. coli metabolism. First, we outline a new computational framework, COBRAme, that empowers the use and facilitates the reconstruction of models of metabolism and gene expression (ME-models). These models offer a comprehensive method to study protein use in microbes and how their metabolism is impacted by the evolutionary pressures to allocate proteome most efficiently. Previously, ME-models were prohibitively difficult to use, but the development of COBRAme has optimized the ME-model reconstruction process making them smaller, easier to understand, and quicker to solve. Second, the next-generation E. coli metabolic model–composing the metabolic core of the ME-model–is detailed. It was further demonstrated that such models can be applied to contextualize multi-omics data types and broaden our understanding of E. coli as a species. Third, adding to the species characterization of E. coli, we leveraged the E. coli ME-model to study how enzyme cofactor availability can shape condition-dependent metabolism. Given that some strains of E. coli are auxotrophic for cofactors, this information provides insight into the consequence and evolutionary drivers of auxotrophy. Lastly, a community E. coli ME-model was constructed to study the adaptation of syntrophy in co-cultures of E. coli auxotrophs. The model provided predictions of how the proteome efficiency of strains in co-culture could affect community characteristics. The totality of this work demonstrates that the scope and applications of these models can be expanded to obtain valuable information about the characteristics of E. coli as a single strain, a species, and in community.