Metabolic simulation using parameters estimated by a distributed genetic algorithm

Abstract

Simulation of the dynamic behavior of the metabolites requires initial amounts of metabolites and all the reaction rate constants to be known. However, it is difficult to obtain actual amount of the metabolites in the experiment from a mass spectrometry, and reaction rate constants are not always known. In the present study, we develop a tool to estimate those parameters that can reproduce the metabolome data obtained from the experiment using a distributed genetic algorithm. We used the parameters to a stochastic simulation, and then reproduced the experimentally observed dynamics of amino acid biosynthesis in Arabidopsis thaliana. Parameters estimated with a distributed genetic algorithm can reproduce more appropriate behavior of the metabolism than that with a genetic algorithm. This method can provide users an estimation of the dynamic behavior of the metabolites whose data are unavailable from the experiment, as well.