ISBN 9783868538502

978-3-86853-850-2, Reihe Informatik

Andreas Dräger
Computational Modeling of Biochemical Networks

254 Seiten, Dissertation Eberhard-Karls-Universität Tübingen (2011), Softcover, A5

Zusammenfassung / Abstract

Systems biology aims to understand mutual interactions and influences between biological components. Control mechanisms and regulation of processes, which often show a nonlinear behavior, are to be described. Predicting the dynamics of complex interaction systems constitutes its main challenge. In this way, it shifts the focus of biology from the mainly descriptive exploration towards a holistic investigation of the complex interactions within living systems.

Computational modeling methods play a crucial role to attain this goal. Physico-chemical, in particular thermodynamic, constraints restrict the investigated systems and define the domain of plausible and valid models. Setting up such a mathematical description is not only complicated and highly error-prone, it also requires knowledge from many different fields and is therefore not easily applicable for non-experts.

This thesis aims at the development of a method that performs the model construction steps to the widest extent automatically, reducing the number of necessary human interactions to a minimum but that still leads to thermodynamically feasible and correct models.

To this end, it introduces a five-step modeling pipeline that ultimately leads to a mathematical description of a biochemical reaction system. We discuss how to automate each individual step and how to put these steps together. The applicability and functioning of these approaches is systematically demonstrated on a model of the valine and leucine biosynthesis in C. glutamicum.

This computer-aided modeling pipeline is further developed to a fully automatic procedure, the AMUSE algorithm (Automated Modeling Using Specialized Enzyme kinetics). Based on latest estimation methods for standard reaction Gibbs energies, AMUSE determines a thermodynamically feasible, minimal equilibrium configuration of the system, identifies the key reactions and selects kinetic equations describing all reaction velocities. Furthermore, it estimates all remaining parameters with respect to given experimental data. The result is an optimized model with respect to its structure and dynamic behavior.