20. March 2014, 14:15
Ernst-Abbe-Platz 2, seminar room 3423

The major difficulty in parameter estimation comes from parameter correlations

Prof. Dr.-Ing. habil. Pu Li
(Simulation and Optimal Processes Group, Institute for Automation and Systems Engineering, TU Ilmenau)

Joint seminar with the Frege Centre for Structural Sciences, Jena

Parameter estimation of dynamic biological models represents a significant challenge in systems biology. This is because, on the one hand, biochemical models usually consist of a large number of parameters to be estimated. On the other hand, in most cases there is no *a priori* knowledge about the parameter values. Therefore, parameter estimation for such models leads to solving a nonlinear, dynamic optimization problem to search for optimal parameter values in a high-dimensional, unconstrained parameter space. The problem will be much more difficult to solve when the parameters are correlated, i.e. mathematically related to each other through some implicit functions, which will lead to the issue of non-identifiablility. In our recent study (Li and Vu, BMC Systems Biology, 7:91) we proposed a method that is able to identify both pairwise parameter correlations and higher order interrelationships among parameters in nonlinear dynamic models. Correlations are interpreted as surfaces in the subspaces of correlated parameters. Based on the correlation information obtained in this way both structural and practical non-identifiability can be clarified. The result of our correlation analysis provides a necessary condition for experimental design in order to acquire suitable measurement data for unique parameter estimation. However, this method can only address *perfect* correlations which are independent of the range of parameter values. In this talk, the extension of our study into the analysis of *imperfect* correlations which depend on the locations in the parameter space will be reported.