The living cell is an autocatalytic network of metabolic pathways sustained far from equilibrium by the supply of matter and energy. At the abstract level we can regard a chart of cellular metabolic pathways as a network of undirected connexions between metabolites (the nodes of the network) each connected pair being related by an enzyme-catalyzed reaction. The clustering properties of the reaction networks can be obtained from maps of known metabolic pathways. For the number of nodes as a function of the number of connexions, a long-tailed distribution is obtained, which can be described by a power law. We also find evidence consistent with a power law in the relationship between regulatory proteins and genes. We investigate three models for the construction of metabolic networks, which we call the random connexion model, the random cluster model and the accumulation model. The last two of these give a long-tailed distribution of nodes. The random cluster and accumulation models also exhibit ‘small-world’ features, in agreement with the structure of real biological networks. We speculate on the possible implications for the evolution of metabolic networks.

Keywords: evolution, genes, metabolism, self-organized criticality, small world networks