The Food Standards Agency GO2 programme into the safety assessment of novel foods provided a major resource of metabolite fingerprint/profile data generated on a range of instruments in different laboratories.

Studies at both SCRI and the University of Aberystwyth showed that differences in data structure confounded statistical analyses aimed at producing a common understanding of metabolite baseline.

The collaborative project, 'Development of unified data models and data pre-processing strategies for the generation of meaningful, standardised statistical analyses of metabolome variability in crop plants', funded as part of the Food Standards Agency G03 programme, will collate and evaluate metabolomic analytical data from all selected G02 contractors and explore methods for alignment and normalisation. It is being led by Derek Stewart, SCRI, Jim McNicol, Biomathematics and Statistics Scotland (BioSS) and John Draper, University of Aberystwyth.

One output will be a validated data pre-processing strategy for production of a standardised 'expected' metabolite peak list bespoke for each crop; for example, in the case of potato the effect of geography and agronomic regime will be assessed.

A robust statistical methodology to generate meaningful metrics for describing metabolite baselines will be validated. Data will be consolidated into a common structure and housed in a bespoke database, with interfaces for selection, filtering and retrieval.