A multinational working group set up to recommend solutions to the daunting problem of integrating the profusion and diversity of metabolomics data has issued a report on its findings.
The Standard Metabolic Reporting Structures working group (SMRS) reported its summary recommendations in the July issue of the journal Nature Biotechnology.
The field of metabolomics blends statistical analysis with state-of-the-art chemical technologies such as nuclear magnetic resonance, mass spectrometry, and chromatography. Applications range from pre-clinical drug safety testing to disease diagnosis to environmental monitoring. But such a diversity of technologies, particularly compared to other “-omics” fields, complicates and intensifies the need for standardization.
The SMRS working group was set up in 2003 with the goal of creating open standards for conducting and reporting metabolomic studies. Contributors include the FDA, the European Bioinformatics Institute, senior academics, and a distinguished group of industry representatives from companies such as GlaxoSmithKline, Pfizer, Unilever, and AstraZeneca.
The Nature Biotechnology article – a synopsis of the SMRS group’s 37-page report – focuses on three main areas: the origins of biological samples, the methodologies used to obtain data, and the analytical techniques to which the data are subjected.
The report offers detailed policy directions in each of these areas. Highlights include guidance on recording the treatment of living biological samples, the naming of compounds in submissions to journals and databases, and how to reduce the problem of over-fitted statistical models and their resultant weaker predictive power.
Lead author John Lindon, from Imperial College London, views the work of the SMRS as “essentially done.” He adds: “The next stage is to take these and build technical and software and database solutions.” The report invites discussion of the recommendations (available at: http://www.smrsgroup.org/documents/SMRS_policy_draft_v2.3.pdf).
The SMRS effort to standardize reporting of metabolomics studies should be viewed as part of a broader, ongoing endeavor in systems biology. Metabolomics will be crucial in illuminating the thousands of pathways in any biological system. Tellingly, SMRS acknowledges the progress made in proteomics, microarray data, and pharmacogenomics. It is “in contact” with similar standardization groups, and expresses a desire to “harmonize with the analogous MIAME initiative in genomics.” This would institute a standard for data content, rather than data format. An FDA-sponsored standard data format has been established, however, with tools under development at California-based PharmQuest.
As a project that seeks to bring into line studies in both public and private domains, SMRS acknowledges that the prickly problem of satisfying both the movement toward openness and commercial sensitivity concerns could become a sticking point. Lindon concedes, “We did not really address IP issues other than to note them and recognize that this might be a problem.”
The question of data ownership in submissions to journals and regulatory bodies is another “major concern” for the group. It seems likely that a pragmatic approach will be adopted, with companies able to preserve confidential information as they see fit, in line with precedents in the field of protein structures, for example. Partial data submissions on a “need to know” basis would furnish the review process.
Further steps in the standardization of metabolomics data will be made at a workshop hosted by the NIH on August 1-2, 2005, in Bethesda, Maryland.
The SMRS report is available here.