| The mathematical modeling of metabolic networks is desirable, but is a difficult task due to a lack of kinetics. To move toward this goal, we have developed an approach, Ensemble Modeling (PM), whereby an ensemble of dynamic models is constructed that reaches the same steady state, is based on an elementary reaction framework, and is constrained by thermodynamics. If the results of experimental phenotypes are known, the models can be similarly perturbed and the ensemble can be screened to retain only models that match the experimental behavior, allowing EM to learn from experiment and bypass the need for a detailed characterization of kinetic parameters.;To validate EM's applicability, the approach was applied to published studies for the production of aromatics in Escherichia coli. In this case study, it was shown that EM can incorporate known experimental data to screen the ensemble and that the retained subset becomes increasingly predictive. Next, to understand the role that the metabolic network structure plays in determining the tendency for certain reactions to be rate-controlling, we utilized the EM framework to randomize the network kinetics and demonstrated that there exists a tendency for certain enzymes to be rate-limiting. In the case of a linear pathway, the first enzyme has this tendency, which is consistent with biological observations. Then, using case studies for aromatics, succinate and lysine production, certain enzymes were found to have an inclination to be rate-limiting, defined as the rate-controlling hotspots, and agree with targets described in the literature. Finally, EM was applied to the metabolism of the nonsulfur purple photosynthetic bacteria Rhodobacter sphaeroides and demonstrated that inactivation of the CO2 assimilatory pathway prevents the system from reaching a redox balance, but can restore growth only through activation of the hydrogen production or sulfate reduction pathway.;The work presented here focuses on the development of a novel technique for the modeling of metabolic networks, the validation of this technique, and its application to novel systems. We believe this approach can significantly enhance one's ability to re-engineer metabolism and provide hypotheses to drive experimentation. |
