A new approach to bioconversion reaction kinetic parameter identification

TLDR

Common bioconversion kinetic parameter identification relies on linear plotting or nonlinear regression, but linear methods demand many experiments and nonlinear regression is prone to local optimization due to dependence on initial values. This study introduces a new, reliable nonlinear regression–based approach for estimating bioconversion kinetic parameters. The method obtains preliminary parameter values stepwise, limiting each step to 3–4 parameters, and then uses these as initial guesses for final nonlinear regression, illustrated with transketolase synthesis of 1,3‑dihydroxypentan‑2‑one. Compared with linear plotting, the approach markedly reduces the number of experiments required for parameter estimation. © 2008 American Institute of Chemical Engineers AIChE J, 2008.

Abstract

Abstract The commonly used methods for bioconversion kinetic parameter identification are linear plotting and nonlinear regression. However, linear plotting methods generally require considerable experimentation, and nonlinear regression can lead to “local optimization” because the obtained parameters depend heavily on given initial values. In this article, a new and reliable nonlinear regression‐based approach to bioconversion kinetic parameter estimation is reported. By obtaining preliminary values of kinetic parameters on a step‐by‐step basis, the number of estimated parameters in each step can be reduced to 3 or 4. These preliminary values can then be used as initial guesses for the final parameter estimation via nonlinear regression. Compared with the linear plotting method, the proposed approach can significantly reduce the number of experiments required for kinetic parameter estimation. The transketolase catalyzed synthesis of 1,3‐dihydroxypentan‐2‐one from propionaldehyde and β‐hydroxypyruvate is used as an experimental example to illustrate the approach. © 2008 American Institute of Chemical Engineers AIChE J, 2008