Dr Eberhard O. Voit, distinguished mathematician and biologist and a world leader in biological systems research and analysis, will give a talk entitled "Metabolic modeling with time series data. Dealing with complexity".
Time: Monday 29 November, 13:15 - 14:00Place: Agricultural University of Norway, Animal Science Building, auditorium H185
Speaker: Eberhard O. Voit, PhD, Medical University of South Carolina, and Wallace H. Coulter Department of Biomedical Engineering - Georgia Tech and Emory University, USA
Dr. Voit's presentation (pdf)
Novel high-throughput techniques produce dense time series of in vivo measurements of the expression of genes at the genomic scale, of many simultaneous concentrations of metabolites, or of the prevalence and activation states of proteins at the proteomic scale. These data will dramatically affect strategies for modeling, analyzing, and optimizing metabolic systems.
At present, the analysis of metabolic systems follows one of two paths. The
first focuses purely on the stoichiometry of the metabolic network and has as its goal the identification, analysis, and manipulation of flux distributions. This approach leads directly to linear node equations, which, in turn, allow a vast spectrum of algebraic and computational tools that can be applied to networks of almost arbitrarily large magnitude. Linearity is also the greatest drawback of the stoichiometric focus, because it is not possible to account for regulatory features, even though they clearly affect the functioning of metabolic networks in a fundamental fashion.
The second path toward understanding metabolic networks is based on nonlinear kinetic descriptions, which are naturally much more flexible than linear systems. This flexibility comes at a much higher cost than for the linear approach, because there are infinitely more nonlinear than linear possibilities for setting up models, which necessitates a non-trivial selection and classification task, and because nonlinear formulations usually require more parameters that must be estimated from experimental data. The design of a nonlinear kinetic model presently consists of four generic steps: (1) identifying which variables and processes are to be considered; (2) characterizing each variable and process based on information from the literature; (3) integrating this information in systems equations; and (4) adjusting parameters secondarily to minimize discrepancies between model responses and observations.
Two crucial challenges of this process are that there are few objective criteria supporting the choice of a particular mathematical process description and that the data necessary for estimation had very often been obtained under different conditions or even from different strains or organisms. The presentation projects how genomic, metabolic, and physiological time series data, combined with novel computational methods, for instance, based on Biochemical Systems Theory, will change the present strategies of nonlinear metabolic modeling so much that it seems legitimate to speak of a new era of metabolic systems analysis.
For more information on Dr Voit, his work and research interests, please see this link.
