Analysis of Large-Scale Metabolic Networks : Organization Theory, Phenotype Prediction and Elementary Flux Patterns
Analyzing metabolic networks is one of the central topics in systems biology. Thus, different methods for analyzing the structure of such networks have been proposed. The methods used here can be broadly divided into two types: ﬂux-centered approaches like elementary mode analysis (EM analysis), the closely related extreme pathway analysis as well as ﬂux balance analysis (FBA) and approaches like chemical organization theory (OT), that additionally explicitly takes into account metabolites. The aim of this work is the integration of these concepts in order to allow for a more comprehensive analysis of metabolic networks. Indeed, EM analysis and OT can complement each other in two ways. First, the set of chemical organizations of a reaction network allows for a clustering of EMs and helps to identify those EMs that cannot operate at steady state. Second, the set of EMs can be used to compute chemical organizations. In another direction, the combination FBA and EM analysis helps to overcome the problem that the entire set of EMs cannot be computed in large metabolic networks. The framework behind this integration, elementary ﬂux pattern analysis (EFP analysis), allows one to identify all possible routes through a subsystem that are part of a pathway in a genome-scale metabolic network. An important benefit of this concept is that it allows to apply many approaches building on EM analysis to genome-scale metabolic networks. Furthermore, using EFP analysis, we identified several EMs in a subsystem of a metabolic model of Escherichia coli that are not compatible with a pathway on the genome scale. Additionally, we discovered several alternative routes to metabolic pathways in the central metabolism of E. coli. Using EFP analysis in a genome-scale metabolic model of humans we found several pathways that contradict the widely held assumption in biochemistry that the conversion of even-chain fatty acids into glucose is infeasible in humans.