TRENDS in Biochemical Sciences Vol.26 No.2 February 2001 http://tibs.trends.com 137 Forum http://tibs.trends.com 0968-0004/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S0968-0004(00)01759-X Forum Metabolic databases – what next? Juliet A. Gerrard, Ashley D. Sparrow and Jessie A. Wells In July 1998, Karp reviewed a new type of bioinformatic resource – metabolic databases 1 . Metabolic databases contain information on collections of reactions or enzymes, or both, which is typically organized into pathways and often linked to genomic information. Such databases are now available on-line, and they provide a convenient means of accessing the ever-increasing morass of biochemical information that is accumulating in the post-genomic era. These data are useful to a wide range of scientists, from those asking fundamental questions about metabolism to those working in applied fields, such as medicine and bioengineering. This article will briefly update the state of the art in metabolic databases, critique the analytical framework in which these databases operate, and then suggest novel approaches for the manipulation of such data, derived from a systems science perspective 2,3 . M etabolic databases in the year 2000 There are many variations on the theme of the on-line metabolic database (Box 1). Some sites contain information that is based on the biochemistry of a single organism. For example, the EcoCyc database claims to contain a complete description of Escherichia coli metabolism. Others, such as ENZYME, contain a more general set of information. There has been a tendency for larger metabolic databases to be regarded automatically as significant new research tools to overcome limitations of empirical biochemistry 4 , but this could be unduly optimistic. Perhaps a change in analytical philosophy is required? All available databases operate within the same paradigm: in terms of both structure and analytical approach, the databases are ultimately electronic versions of the metabolic map that has adorned the walls of biochemistry laboratories for 50 years. The electronic nature of the latest metabolic databases facilitates links to genomic information, its expression and regulation. They also offer sophisticated ways in which to search the metabolic map, ask specific questions and explore metabolic routes not immediately apparent on the hard copy. However, the databases remain, Forum Computer Corner Aspartate 4-Aspartyl phosphate Aspartyl semi- aldehyde Homoserine O-Phospho- serine Threonine 2-Keto- butyrate 2-Thiamine Enzyme a Enzyme b Enzyme c Enzyme d Enzyme e Enzyme g Enzyme f 2,3-Dihydro- dipicolinate Tetrahydro- dipicolinate N-succinyl 2- amino-6-keto- pimelate N-succinyl 2,6-diamino- pimelate L,L-diamino pimelate Meso-diamino pimelate Enzyme h Enzyme i Enzyme j Enzyme k Enzyme l Enzyme n Enzyme m O-Succinyl homoserine Cysta- thionine Homo- cysteine Enzyme o Enzyme p Enzyme r Methionine Enzyme s 2-Aceto, 2-hydroxy- butyrate 2,3-Dihydroxy- butyrate 2-Keto-3- methylvalerate Enzyme t Enzyme u Enzyme v Isoleucine Pyruvate Enzyme q Lysine Enzyme f Enzyme s Enzyme q Ti BS Fig. 1. The traditional view of the biosynthesis of the aspartate family of amino acids, adapted from the Boehringer–Mannheim map (obtained from Boehringer–Mannheim, Mannheim, Germany). In this metabolite-centric view of the cell, emphasis is given to the metabolites rather than the macromolecules that catalyse their interconversion. Enzymes a–v are as follows: a, aspartokinase; b, aspartyl semi-aldehyde dehydrogenase; c, homoserine dehydrogenase; d, homoserine kinase; e, threonine synthase; f, dihydrodipicolinate synthase; g, dihydrodipicolinate reductase; h, tetrahydrodipicolinate succinylase; i, succinyl-dap transaminase; j, succinyl-dap desuccinylase; k, dap epimerase; l, dap decarboxylase; m, homoserine acyl transferase; n, cystathionine synthase; o, cystathionine-β-lyase; p, homocysteine methyl transferase; q, cystathione-γ-lyase; r, threonine deaminase; s, acetolactate synthase; t, dihydroxyacid dehydratase; u, acetohydroxyacid isomerase; v, glutamate-dependent transaminase.