(391a) Evaluating Metabolic Network Elements Via a Forced Coupling Algorithm

Reed, J. L., University of Wisconsin-Madison

The development of genome-scale models (GEMS) has allowed for the interrogation of microbial reaction networks for a variety of purposes such as predicting growth phenotypes, guiding metabolic engineering strategies, elucidating network properties and facilitating hypothesis-driven discovery [1]. However, despite the many applications of GEMS, there remains room for improvement and development, both in terms of the networks themselves, as well as the algorithms used to interrogate such systems. The success of in silico cellular models depends on the ability of both the metabolic and regulatory networks to capture the underlying characteristics of the microbe of interest. Unfortunately, the introduction of error into these networks during construction is often unavoidable due to errors in databases or functional annotations, or to reactions added during pathway refinement.

To facilitate model testing and validation, we have developed a forced coupling algorithm (FCA) that can design experiments which test for the existence of network elements, whether these elements are hypothesized reactions or gene to reaction associations. The FCA makes use of the concept of flux coupling [2] to propose mutant strains and media conditions under which a particular network element is conditionally essential for cell growth or coupled to another measurable reaction (e.g. chemical production). By testing such conditions experimentally, hypothesized or questionable network elements can be confirmed or dismissed and network models can be further refined to better represent the true physiology of the microbial organism.

This algorithm has been applied to a variety of genome-scale models, including Escherichia coli (iJR904) [3], Bacillus subtillis (iYO844) [4] and Pseudomonae putida [5], and has been successful in coupling over 80 percent of their accessible metabolic network to biomass production. For E. coli, the FCA results were compared to existing experimental datasets [6, 7] to confirm or refute gene to reaction relationships based on measured gene essentiality. For reactions that were coupled to biomass using only media coupling conditions (without requiring any gene deletions) the experimental results supported the FCA predictions (in ~80% of cases), indicating that these reactions are used under these particular conditions.

The FCA coupling conditions can be used to validate existing cellular network annotations such as models produced using the Model SEED architecture.[8] Moreover, these conditions can be used as screens to identify genes responsible for orphan reactions (reactions whose associated genes are unknown), thus filling important knowledge gaps in metabolism.  Additionally, we have integrated the FCA algorithm into metabolic strain design algorithms to propose unique E. coli mutants that if evolved would co-utilize multiple carbon substrates and produce ethanol at high yields. The FCA algorithm can be used for a variety of potential applications (model testing, experimental design, and metabolic engineering) and these will be presented.


CT is supported by an NHGRI training grant to the Genomic Sciences Training Program (5T32HG002760).


1.            Oberhardt, M.A., B.O. Palsson, and J.A. Papin, Applications of genome-scale metabolic reconstructions. Mol Syst Biol, 2009. 5.

2.            Burgard, A.P., et al., Flux Coupling Analysis of Genome-Scale Metabolic Network Reconstructions. Genome Research, 2004. 14(2): p. 301-312.

3.            Reed, J., et al., An expanded genome-scale model of Escherichia coli K-12 (iJR904 GSM/GPR). Genome Biology, 2003. 4(9): p. R54.

4.            Oh, Y.-K., et al., Genome-scale Reconstruction of Metabolic Network in Bacillus subtilis Based on High-throughput Phenotyping and Gene Essentiality Data. Journal of Biological Chemistry, 2007. 282(39): p. 28791-28799.

5.            Oberhardt, M.A., et al., Reconciliation of Genome-Scale Metabolic Reconstructions for Comparative Systems Analysis. PLoS Comput Biol, 2011. Accepted.

6.            Baba, T., et al., Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol, 2006. 2.

7.            Joyce, A.R., et al., Experimental and Computational Assessment of Conditionally Essential Genes in Escherichia coli. The Journal of Bacteriology, 2006. 188(23): p. 8259-8271.

8.            Henry, C.S., et al., High-throughput generation, optimization and analysis of genome-scale metabolic models. Nat Biotech, 2010. 28(9): p. 977-982.