Computationally predicting rate constants in pathway models

Peter Henning; Richard Moffitt; Jeremy Allegood; Elaine Wang; Alfred Merrill; May Wang

doi:10.1109/IEMBS.2005.1615622

Computationally predicting rate constants in pathway models

Conf Proc IEEE Eng Med Biol Soc. 2005:2005:5093-6. doi: 10.1109/IEMBS.2005.1615622.

Authors

Peter Henning¹, Richard Moffitt, Jeremy Allegood, Elaine Wang, Alfred Merrill, May Wang

Affiliation

¹ WHC Department of Biomedical Engineering, Georgia Institute of Technology and Emory University.

PMID: 17281392
DOI: 10.1109/IEMBS.2005.1615622

Abstract

The purpose of this project is to elucidate the kinetic parameters that govern a simulated sphingolipid metabolism system using various global optimization routines including Monte Carlo, Simulated Annealing, and Genetic Algorithms. Here a simulated 6-node system built from five UniUni reaction equations with known kinetic parameters. Each node was treated as a combination of single substrate - single product catalyzed reactions. This defined system of equations is then sampled at a rate that mimics the mass spectrometry measurements of the complex pathway in time shown in [1]. As the investigation on mathematical models of biological events continues to gain popularity, the use of global optimization methods to quickly and reliably estimate missing parameters will become more vital. This work investigates the use of global parameter estimation schemes in terms of their reliability to the true underlying kinetic parameters. When the amount of fitting parameters is sufficiently large, it is likely to find parameter sets that predict the data decently well but are not necessarily close to the true underlying parameters.