A new algorithm for the kinetic data analysis

In this paper, a new algorithm for the kinetic data analysis is presented. The main objective of the algorithm is to retrieve the maximum information concerned with a multi-response complex chemical system evolving in time, i n order to retrieve the rate constants (calibration problem) or the initial concentration of species. As a difference with other data treatments found in the literature, the algorithm is able to estimate the uniqueness and re liability of the calculated rate constants. This task is carried out by ana lyzing of the principal components of the sensitivity coefficients with reg ard to the rate constants. The analysis allows understanding whether the lo cated stationary points consist of a single point, or a surface relating a set of rate constants on which the least-squares (LS) function takes a cons tant value. In the latter case, the rate constants will not be uniquely det ermined for the chosen experimental design. The algorithm has been materialized in the OPKINE 2 software, which has bee n provided with a wide range of mathematical resources related to the minim a location, and to the integration of stiff ordinary differential equations (ODE) systems. Finally, the statistical criteria used to compare the exper imental data with a model are not only restricted to the LS method, and it is possible to use the determinant criterion, or to fit the principal compo nents of the experimental responses. The application of the algorithm to fi rst-order networks allows saving up a great quantity of calculation time co mpared to classical numerical integration. (C) 2000 Elsevier Science B.V. A ll rights reserved.