Search

The 1st international workshop on the ‘Food metabolome and biomarkers for dietary exposure’, organized in Glasgow last year, identified as a priority the development of databases and libraries of spectra for the food metabolome. The food metabolome comprises all metabolites present in human biofluids and tissues that directly derive from the digestion and metabolism of food chemicals. Exploration of the food metabolome through mass spectrometry untargeted profiling has opened new avenues for the discovery of intake biomarkers, also a key priority in nutrition research as described in the JPI HDHL strategic agenda and recent BioNH call. A large proportion of the food metabolome consists of metabolites of phytochemicals such as polyphenols and terpenes. Identification of these metabolites in metabolomic profiles is often a bottleneck in biomarker discovery, as their standards are not commercially available and their integration in online databases and libraries of spectra is still very limited. As part of the ANR PhenoMeNep project, we have designed PhytoHub (www.phytohub.eu), dedicated to all phytochemicals commonly ingested with the diet. Chemical data include name, chemical structure and identifiers, synonyms, physico- chemical properties. Data also include the taxonomy, foods of origin (linked to FooDB), compound of origin (if a metabolite), biofluid location and literature references. Known metabolites are manually extracted from the literature and references are attached. Since the metabolism of many phytochemicals has not been studied yet, in silico prediction of metabolism will be used. We are developing an in-house tool from (1) a compilation of all biotransformations (and combinations) occurring in humans (including gut microbiota biotransformations) for each chemical class and (2) analysis of functional groups on precursor phytochemicals. A list of predicted metabolites will be generated for each phytochemical. Mass spectral data come from literature, other databases on phytochemicals and experimental data from our collaborative platforms. Hopefully in silico prediction of mass fragmentation will be added in the future. An efficient relational design supports a powerful and intuitive web interface. For a queried monoisotopic mass or molecular formula, the database will return a list of metabolites or phytochemical precursors, along with their spectral data and possible dietary and metabolic origins. For a queried food, it will return a list of metabolites likely to be present in biofluids after consumption. MS-MS and custom advanced searches are also possible. PhytoHub is the first database to collate information on phytochemical metabolites from a metabolomics standpoint, and should facilitate identification of unknowns in non-targeted profiling. The version 1.0 of PhytoHub is online, with a first dataset on the 240 most consumed terpenes and their known metabolites. It will be updated monthly. Any willingness to contribute to this freely accessible resource is welcome so that the database can be as complete, accurate and useful as possible.