MetaDAG: a web tool to generate and analysemetabolic networks

Abstract

We introduce MetaDAG, a web-based tool designed for metabolic network reconstruction and analysis. MetaDAG is capable of constructing metabolic networks associated with specific organisms, sets of organisms, sets of reactions, sets of enzymes, and sets of KO (KEGG Orthology) identifiers. To generate these metabolic networks, MetaDAG retrieves from the KEGG database the chemical reaction information that corresponds to the users queries. MetaDAG computes a reaction graph as a first metabolic graph model. This reaction graph is a network in which nodes represent reactions, and edges between reactions indicate the presence of a metabolite produced by one reaction and consumed by another. Next, as a second metabolic model, MetaDAG computes a directed acyclic graph called a metabolic DAG (m-DAG for short). The m-DAG is obtained from the reaction graph by collapsing all strongly connected components into single nodes. As a result, the m-DAG representation reduces considerably the number of nodes while keeping and also highlighting the networks connectivity. Both metabolic models, the reaction graph, and the m-DAGs, are displayed on an interactive web page to assist the users in visualising and analysing the networks. Furthermore, users can retrieve the nodes information linked to the KEGG database. All generated files, including images containing metabolic network information and analysis results, are available for download directly from the web page. In the Eukariotes test presented here, MetaDAG has demonstrated its effectiveness in classifying all eukaryotes from the KEGG database at both the kingdom and phyla taxonomy levels.