Skip to main content - access key m.
Skip to main navigation - access key n.

Friday, March 25, 2011
10:30 a.m., ECSS 2.415












“Stoichiometric Modeling of Biochemistry: A Variational Approach”
Dr. Ronan M.T. Fleming, University of Iceland

In molecular systems biology, genome scale models are inherently high dimensional. With such models, one’s choice of modeling approach is often swayed by concerns of computational tractability rather than strict adherence to underling physicochemical principals. One prominent example of this is the widely used genome scale modeling approach termed flux balance analysis, which relies heavily on linear programming. After an introduction to stoichiometric modeling we shall describe recent advances that allow for increased physicochemical realism yet maintain tractability. These advances rely on a novel blend of parametric convex optimization and Lyapunov stability theory.

The application of such theory is vital for developing algorithms that can be guaranteed to converge to a physicochemically realistic solution, assuming it exists. A solution may be interpreted as a stable steady-state flux and concentration vector for a non-equilibrium thermodynamic system. More generally, we describe related open problems in genome scale modeling of biochemical systems and argue that convex optimization will play a central role in solving such problems.

Ronan M.T. Fleming trained as a veterinarian and histo-pathologist prior to focusing on mathematical and numerical modeling in systems biology in 2003. A group leader at the Center for Systems Biology at the University of Iceland, his focus is on improving the predictive capacity of constraint-based modeling by rigorously incorporating non-equilibrium steady-state thermodynamics into the in silico represention of biochemical networks. The mathematics of convex optimization is a particular interest since it lies at the very foundation of a comprehensive variational approach to modeling biochemical networks in a far-from-equilibrium steady state. The aim is apply these fundamental modeling advances to the study of neoplasitc cell metabolism.