Data Description and Discoverability

Making data FAIR requires description at two levels: the experiment and the repository. At the experiment level, OBI provides a formal vocabulary for study design and protocols. NIDM represents the full neuroimaging research workflow as a machine-readable provenance graph built on PROV-O. UBERON provides cross-species anatomical annotation used across repositories and standards for brain region labelling. At the repository and infrastructure level, DCAT is the W3C catalogue vocabulary underpinning dataset discoverability across EOSC and Recherche Data Gouv. Dublin Core provides a widely adopted base metadata layer across repositories. Persistent identifiers complete the picture: ROR identifies institutions, ORCID identifies researchers, RRID identifies research resources and core facilities, and DataCite assigns DOIs to datasets and software through its Fabrica registration service, with the DataCite Metadata Schema defining how those objects are described. Zenodo and Figshare are the primary generalist repositories built on this DOI infrastructure, each accepting any file type and making any research output (datasets, figures, code, preprints, and presentations) persistently citable and openly accessible.