COARSE BUT EFFICIENT IDENTIFICATION OF METABOLIC PATHWAY SYSTEMS

The rapid development of high-throughput experiments in molecular biology over the past decade has facilitated the generation of metabolic datasets of unprecedented quantity and quality. These datasets contain enormous information that must be extracted with computa- tional methods. A premier goal of such an extraction is the construction of fully parameterized kinetic models of metabolic pathway systems, which may subsequently be utilized for deepening our understanding of metabolism and for the design of manipulation and optimization strat- egies toward the production of valuable organics. Numerous methods have been proposed for the conversion of metabolic data into such models, but the nonlinear nature of regulated pathway systems continues to be a challenge for the estimation of reliable model parameters. To alleviate the situation, the present work proposes a coarse yet efficient technique for the estimation of parameter values that is based on linear regression and therefore naturally scales to large systems. The proposed method permits the complete construction of a coarse model, the identification of a starting point for fine-tuned nonlinear estimation, or the characterization of a few fluxes that can subsequently be used for performing Dynamic Flux Estimation (DFE), a method that yields insights into the true nature of the processes that govern metabolic path- way systems. The proposed linear inference method is demonstrated with a regulated branched pathway and with a metabolic system from the literature that describes the biosynthesis of amino acids from aspartate. Both analyses indicate that the method yields satisfactory results . Keywords- Biochemical Systems Theory (BST), Dynamic Flux Estimation (DFE), Metabolic pathway analysis, Parameter estimation, S- system, Time course data