Physical understanding via reduction of complex multiscale models: Glycolysis in saccharomyces cerevisiae

We consider complex mathematical models that are characterized by a wide spectrum of time scales, the fastest of which are operative during the initial state only, leaving the slower ones to drive the system at later times. It is shown that very useful physical understanding can be acquired if the fast and slow dynamics are first separated and then analyzed. Existing algorithmic methodologies can be applied for this purpose. A demonstration of this approach is presented for a glycolysis model, the solution of which asymptotically evolves around a limit cycle.