Metabolomics and the challenges ahead

Eight years have passed since the proposal that the characterization of mutations with silent phenotypes should be carried out by a global view of metabolism [1]. Since then, the field of metabolomics has flourished both in the development of new techniques and in the expansion of its application areas. No longer is metabolomics seen just as a tool of functional genomics, but it has now become an integral part of systems biology. Early approaches were based on metabolite fingerprints and metabolite profiles, but now there are also studies focusing on fluxes. The array of techniques to measure a large number of metabolites has also been expanding, with a clear predominance of mass spectrometry and nuclear magnetic resonance (NMR) methods but with new methods based on probes in development. But anyone who has been initiated in this field will recognize that the applications of metabolomics are currently limited by computational issues. The bigger bioinformatics challenges posed by metabolomics are: (1) to identify the large number of metabolites that are detected but whose chemical nature is unknown (estimates go from 60 to 90% of the total); (2) to identify the active areas of metabolism (pathways and networks) responsible for changes in metabolite profiles and (3) to create standards for data and metadata format and reporting. This issue of Briefings in Bioinformatics carries a series of articles dedicated to addressing the bioinformatics challenges posed by metabolomics. To introduce the problem, a review by Shulaev describes the most popular metabolomics techniques and enumerates the bioinformatic issues at that level. Two reviews address the issue of pathways and networks: Lee et al. describe the relationship between flux balance analysis and metabolomics, and Steuer and Selbig cover the many interesting aspects of metabolite correlations and how they may be related with metabolic networks and regulation. Finally, an article by Castle et al. discusses the new ongoing efforts of an international and interdisciplinary group that is defining standards for metabolomics. This follows several earlier attempts at defining standards [2–4] and unites those efforts. The much anticipated result of this effort will be a set of data standards that cover all of metabolomics and will be compatible with other standards for genomics and proteomics. It is the simultaneous application of metabolomics, proteomics, and transcriptomics that is likely to provide the most comprehensive and informative views of biological systems, and for this to happen it is crucial that we have data standards in place. The next step will be to integrate these data in global models of cellular behavior—indeed a truly systems view of biology!