(556g) Qualitative And Quantitative Experimental Design For An Enzymatic Reaction Network

The increasing interest in using biochemical synthesis routes for producing complex fine chemicals and intermediates in the pharmaceutical industry constitutes the general motivation behind the EUROBIOSYN project. Large reaction networks are required to develop a purely enzymatic synthesis for complex molecules from simple (sugar) substrates. One way to construct such a functional enzymatic reaction network is called a System of Bio-transformations (SBT) and is based on a selected part of one single organism's metabolic network containing the synthesis paths including cofactor regeneration reactions. Suitably modified genetic mutants of E-coli microorganism are used in this work to produce the metabolic network for SBT, which is performed as cell free extract in the production phase. The key product is Di-hydroxy-acetone phosphate (DHAP), and the DHAP-producing SBT contains all the enzymes for the glycolysis reactions, leading to a system of high complexity. In order to understand the system functionality and to optimize the production it is desirable to develop quantitative dynamic process models, which exhibit good long-term prediction properties over a wide range of the operating region. During model development one of the important steps is to determine which parameters can be estimated from the available experimental data. A subsequent problem is to determine which model states should be measured and which inputs should be perturbed in order to render all the parameters identifiable. This operation is known in the literature as qualitative experimental design. The procedure is initiated with a reduced set of measured state and possible one perturbed input. A two-step analysis is used to determine whether the model parameters can be estimated for a given set of measured states and perturbed inputs. The first stage aims at finding the identifiable reaction rates (fluxes) based on stoechiometry. In the second stage a method based on nonlinear algebra is used to assess the kinetic parameters identifiability for each of the identifiable rates. The method employs the calculation of Lie derivatives and solution of a system of algebraic equations. The parameter identifiability properties are being assessed based on the number of solutions of these systems of equations. Successively a larger set of measured states and/or perturbed inputs is considered and the procedure repeated until all the parameters are rendered to be identifiable or until the entire model states are considered. Once it has been determined which inputs are to be perturbed and which outputs should be measured, quantitative experimental design is to be performed in order to maximize the accuracy of the parameters estimates. A criterion based on the observed Fisher information matrix is to be used to determine the input perturbations, initial conditions and the time to measure the states. The presentation will focus on experimental design for a model for an enzymatic reaction network for DHAP production.