(295e) Optimization Driven Identification of Genetic Perturbations Accelerating the Convergence of Model Parameters in Ensemble Modeling of Metabolic Networks

The recently introduced ensemble modeling (EM) approach has been proven successful in bypassing the need for experimental determination of individual kinetic parameters and dynamic modeling of metabolic networks. The essence of the EM procedure involves screening the models in the ensemble by virtue of simulating genetic perturbations (i.e., knock out/down/up) resulting in a new steady state flux distribution compared to the wild-type strain. A key element affecting the success of the EM procedure is thus the ability to converge to a set of parameters that best describe the observed phenotype of the system. However, as the set of possible kinetic parameters becomes smaller following some initial screening, subsequent genetic perturbations must become more targeted. In this study we propose an optimization-based algorithm for pro-active identification of genetic/enzyme perturbations that maximally reduce the number of retained models in the ensemble after each round of model screening. The key premise here is to design perturbations that will maximally scatter the predicted steady-state fluxes over the ensemble parameterizations thus allowing for maximal elimination of models from the ensemble and faster convergence to the correct one.  We demonstrate the screening ability of this procedure using a metabolic model of the central metabolism of E. coli and by successively identifying single, double, triple and higher order enzyme perturbations (i.e., knockouts, overexpressions or combinations thereof) that would ensure the maximum degree of flux separation between the models in the ensemble. Overall, our study provides an optimal way of designing genetic perturbations for populating the ensemble of models with relevant model parameterizations.