Parametric sensitivity of stoichiometric flux balance models applied to wild-type Escherichia coli metabolism.

Stoichiometrically based flux balance models provide a method to quantify the metabolic pathway fluxes within a living cell. Predictions of flux balance models are expected to have applications in pathway engineering as well as in bioprocess design and control. These models utilize optimality principles applied to metabolic pathway stoichiometry along with the metabolic requirements for growth to determine the flux distribution in a metabolic network. A flux balance model has been developed for Escherichia coli W3110 using five experimentally determined strain-specific parameters. In this report, we determine the sensitivity of the predictions of the flux balance model to these five strain-specific parameters. Model predictions are shown to be sensitive to the two parameters describing metabolic capacity, while they are relatively insensitive to the three parameters that describe the metabolic requirements for growth. Thus, when stoichiometrically based models are formulated for additional strains one needs to measure the metabolic capacity (maximum rates of nutrient and oxygen utilization) accurately. Determination of metabolic capacity from batch experiments is relatively easy to perform. On the other hand, the harder to determine maintenance parameters need not be as accurately determined. (c) 1995 John Wiley & Sons, Inc.