Sharing data as machine-actionable objects using RO-Crate, Bioschemas and Signposting

RO-Crate (https://doi.org/10.3233/DS-210053) is a lightweight method to package research outputs along with their metadata, based on existing Linked Data standards. Bioschemas (https://bioschemas.org/) provides metadata schemas to add structured metadata to webpages on Life Science, based on schema.org. Signposting (https://signposting.org/) specifies a lightweight yet powerful approach to increase the FAIRness of scholarly objects, based on Web standards. 

The combination of these standard-based approaches make resources easy to navigate by machines, provide an unambiguous way for machines to access FAIR metadata and content in a single request, and reduce content-negotiation hassle that can give unpredictable results. 

This tripartite combination is of benefit for repositories and publishers, as they can non-disruptively add FAIR Signposting headers for machine navigation, support RO-Crate imports and align with Bioschemas specifications, making FAIR Digital Objects achievable with existing technologies over HTTP. 

Our broad communities have a strong foundation in ELIXIR, through individual engagement, as well as in ELIXIR-aligned projects (EOSC-Life, BY-COVID, EOSC4Cancer, EuroScienceGateway, Biodiversity Genomics Europe). Bioschemas is deployed by 67 ELIXIR resources, and its community have built rich profiles and matured biochemistry types like Protein into schema.org.

While ELIXIR’s role in RO-Crate initially was fairly workflow-focused (e.g. WorkflowHub, WfExS), the rest of the RO-Crate community is quite data-centric, with recent adaptation by biodiversity and plant science initiatives. We are therefore now arranging several training events (including at ELIXIR AHM) and virtual training modules to improve uptake of FAIR using our pragmatic approach to rich metadata using established Web technologies.