(418d) Dynamic Metabolic Flux Analysis with Linear Flux Functionality

Metabolic flux analysis (MFA) using stable isotopes has been established as a practical tool for developing flux mappings of complex biological reaction networks. At the present time, work involving MFA has been limited to systems at metabolic steady state. In order to develop a framework capable of analyzing fully dynamic systems, a reaction rate functionality must be assumed. In the current work the validity of approximating reaction rates as a linear functionality of time was examined using principles of stoichiometric metabolic flux analysis. A model was developed using linear variations in fluxes to analytically simulate concentration profiles from initial concentration and fluxes at discrete time points. Weighted least squared principles were applied to estimate optimum values for these flux parameters from concentration and flux measurements in order to elucidate flux profiles. An error minimization routine was applied to select optimum discrete time points for fluctuations in reaction rate in order to minimize error while preventing overfitting. The model was applied to a previously collected data set, concentration profiles for production of 1,3-propanediol obtained from a fed-batch fermentation of a high yielding strain of E. coli. The new model was demonstrated successfully and flux profiles generated were in agreement with flux estimates by traditional means. In addition, a statistical analysis of the model was conducted allowing, for the first time, determination of flux confidence intervals and parameter sensitivities. Establishment of parameter sensitivity to measurements will allow the model to serve as a guide for sample time point selection. Having demonstrated the validity of the linear flux mappings approximation, our current work involves application of linear flux mapping for analysis of dynamic isotope labeling experiments. Additional work will also involve extension of the methods within the current work to analyze the validity of more complex flux mapping functionalities.