Sensitivity analysis and reduction of a dynamic model of a bioproduction of fructo-oligosaccharides

Abstract

Starting from a relatively detailed model of a bioprocess producing fructo-oligosaccharides, a set of experimental data collected in batch and fed-batch experiments is exploited to estimate the unknown model parameters. The original model includes the growth of the fungus Aureobasidium pullulans which produces the enzymes responsible for the hydrolysis and transfructosylation reactions, and as such contains 25 kinetic parameters and 16 pseudo-stoichiometric coefficients, which are not uniquely identifiable with the data at hand. The aim of this study is, therefore, to show how sensitivity analysis and quantitative indicators based on the Fisher information matrix can be used to reduce the detailed model to a practically identifiable model. Parametric sensitivity analysis can indeed be used to progressively simplify the model to a representation involving 15 kinetic parameters and 8 pseudo-stoichiometric coefficients. The reduced model provides satisfactory prediction and can be convincingly cross validated.

A multiscale modelling approach for Haematococcus pluvialis cultivation under different environmental conditions

“…A concatenated matrix containing all the normalised sensitivities is constructed as shown in previous research [ 40 ]. The matrix has the same number of columns as the number of parameters.…”