The MetaboX library: Building metabolic networks from KEGG database

In many key applications of metabolomics, such as toxicology or nutrigenomics, it is of interest to profile and detect changes in metabolic processes, usually represented in the form of pathways. As an alternative, a broader point of view would enable investigators to better understand the relations between entities that exist in different processes. Therefore, relating a possible perturbation to several known processes represents a new approach to this field of study. We propose to use a network representation of metabolism in terms of reactants, enzymes and metabolites. To model these systems, it is possible to describe both reactions and relations among enzymes and metabolites. In this way, analysis of the impact of changes in some metabolites or enzymes on different processes are easier to understand, detect and predict. Results. We release the MetaboX library, an open source PHP framework for developing metabolic networks from a set of compounds. This library uses data stored in the Kyoto Encyclopedia for Genes and Genomes (KEGG) database using its RESTful Application Programming Interfaces (APIs), and methods to enhance manipulation of the information retrieved from the KEGG webservice. The MetaboX library includes methods to extract information about a resource of interest (e.g. metabolite, reaction and/or enzyme) and to build reactants network, bipartite enzyme-metabolite and unipartite enzyme networks. These networks can be exported in different formats for data visualization with standard tools. As a case study, the networks built from a subset of the Glycolysis pathway are described and discussed. Conclusions. The advantages of using such a library imply the ability to model complex systems with few starting information represented by a collection of metabolites KEGG IDs.