Assessment of effects of experimental imprecision on optimized biochemicalsystems

Metabolic pathways may be optimized with S-system models that prescribe pro files of control variables leading to optimal output while keeping metaboli tes and enzyme activities within predefined ranges. Monte Carlo simulations show how much the yield and the corresponding metabolite concentrations wo uld be affected by inaccuracies in the experimental implementation of the p rescribed profiles. For a recent model of citric acid production in Aspergi llus niger, the yield is roughly normally distributed, whereas the distribu tions of metabolite concentrations differ greatly in shape and statistical characteristics. Even moderate inaccuracies may lead to constraint violatio ns, which appear to be correlated with high logarithmic gains. (C) 2001 Joh n Wiley & Sons, Inc.