Characterization Measurements of Metabolic Hypergraphs

Metabolic networks are probably among the most challenging and promising biological networks. Their study provides insight into how biological pathways work and how robust a specific organism is against an environment or therapy. Previous studies have obtained relevant results using flux balance analysis (FBA) and simulations of single gene deletion. However, the structural characterization of metabolic networks as complex networks has been proven an arduous task. Past attempts have considered graphs whose nodes are the metabolites or reactions of the metabolic network in question, and only recently has the focus shifted to higher-order structures, highlighting that simple pairwise interaction may not be sufficient for characterization. Here we show an intuitive way to map metabolic networks into hypergraphs using the bipartite representation. We introduce structural characterization measurements to analyze the metabolism of a single organism and to compare the robustness and complexity of different metabolic models. Our findings show a connection between topological and biological properties. Communicability and information-based measurements succeed in identifying relevant metabolites and reactions in the metabolic hypergraph, and different organisms exhibit an overall similar complexity but very different robustness. In particular, the Staphylococcus Aureus, an antibiotic-resistant bacterium, displays the highest robustness.