Combining graph and flux-based structures to decipher phenotypic essential metabolites within metabolic networks.

Background: The emergence of functions in biological systems is a long-standing issue that can now be addressed at the cell level with the emergence of high throughput technologies for genome sequencing and phenotyping. The reconstruction of complete metabolic networks for various organisms is a key outcome of the analysis of these data, giving access to a global view of cell functioning. The analysis of metabolic networks may be carried out by simply considering the architecture of the reaction network or by taking into account the stoichiometry of reactions. In both approaches, this analysis is generally centered on the outcome of the network and considers all metabolic compounds to be equivalent in this respect. As in the case of genes and reactions, about which the concept of essentiality has been developed, it seems, however, that some metabolites play crucial roles in system responses, due to the cell structure or the internal wiring of the metabolic network.

Results: We propose a classification of metabolic compounds according to their capacity to influence the activation of targeted functions (generally the growth phenotype) in a cell. We generalize the concept of essentiality to metabolites and introduce the concept of the (PEM) which influences the growth phenotype according to sustainability, producibility or optimal-efficiency criteria. We have developed and made available a tool, , which implements a method combining graph-based and flux-based analysis, two approaches that are usually considered separately. The identification of PEMs is made effective by using a logical programming approach.

Conclusion: The exhaustive study of phenotypic essential metabolites in six genome-scale metabolic models suggests that the combination and the comparison of graph, stoichiometry and optimal flux-based criteria allows some features of the metabolic network functionality to be deciphered by focusing on a small number of compounds. By considering the best combination of both graph-based and flux-based techniques, the python package advocates for a broader use of these compounds both to facilitate network curation and to promote a precise understanding of metabolic phenotype.