EFMEvolver: Computing elementary flux modes in genome-scale metabolic networks

Abstract

Elementary flux mode analysis (EFM analysis) is an important method in the study of biochemical pathways. However, the computation of EFMs is limited to small and medium size metabolic networks due to a combinatorial explosion in their number in larger networks. Additionally, the existing tools to compute EFMs require to enumerate all EFMs before selecting those of interest. The method presented here extends EFM analysis to genome-scale models. Instead of computing the entire set of EFMs an optimization problem is used to determine a single EFM. Coupled with a genetic algorithm (GA) this allows to explore the solution space and determine specific EFMs of interest. Applied to a network in which the set of EFMs is known our method was able to find all EFMs in two cases and in another case almost the entire set before aborted. Furthermore, we determined the parts of three metabolic networks that can be used to produce particular amino acids and found that these parts correspond to significant portions of the entire networks. Availability: Source code and an executable are available upon request.