Creating lightweight FAIR Digital Objects with RO-Crate

RO-Crate (Soiland-Reyes et al. 2022) is a lightweight method to package research outputs along with their metadata, based on Linked Data principles (Bizer 2009) and W3C standards. provides flexible mechanism for researchers archiving publishing rich data packages (or any other outcome) by capturing dependencies context. However, additional measures should be taken ensure that crate also following the FAIR (Wilkinson 2016), including consistent use of persistent identifiers, provenance, community standards, clear machine/human-readable licensing metadata data, Web publication RO-Crates. The Digital Object (FDO) approach (De Smedt 2020) gives set recommendations aims improve findability, accessibility, interoperability reproducibility digital object, allowing implementation through different protocols or Here we present how have followed FDO turned outcomes into FDOs RO-Crates using HTTP, best practices Data. We highlight challenges advantages approach, reflect what required an profile achieve allows broad range cases, across scientific domains. A minimal may represented as URI resolving summary website describing in investigation (e.g. https://w3id.org/dgarijo/ro/sepln2022 links used datasets software). One flexibility, particularly regarding accompanying actual outcome. extends schema.org, popular vocabulary resources (Guha 2016). generic not typed beyond Dataset *1. In practice, declare conformance particular profiles, processing specific needs assumptions usage scenario. This, effectively, makes thus machine-actionable. profiles serve templates, making it easier communities agree build upon own needs. been combined machine-actionable Management Plans (maDMPs) automate facilitate management (Miksa 2020). This mapping generated out maDMPs vice versa. ELIXIR Software (Alves 2021) planning move questionnaire format RO-Crate. Biohackathon 2022 will explore integration Stewardship Wizard (Pergl 2019) Galaxy, which can creation follows plans. tailored has defined represent Electronic Lab Notebooks (ELN) bundled together related datasets. Schröder (2022) uses encode provenance information at levels, researchers, manufacturers, biological chemical resources, activities, measurements, resulting data. programmatically question-answer protocols, instance equipment create Another example WorkflowHub (Goble defines Workflow (Bacall 2022), imposing constraints such presence main workflow license. It specifies entity types properties must provide information, implicitly defining operations (e.g., get its language) are valid all complying crates. system Galaxy (The Community retrieves Crates GA4GH TRS API. further extended (with OOP-like inheritance) Testing RO-Crate, adding formal testing components: this adds getting remote test instances definitions, LifeMonitor service keep track health status multiple published workflows. While technologies, they self-contained , moving metadata. powerful construct repositories, but raises some regards mutability persistence To illustrate handled, detail repository several principles: entries frozen editing complete kernel (title, authors, license, description) [FDOF4] before assigned identifier, e.g. https://doi.org/10.48546/workflowhub.workflow.255.1 [FDOF1] Computational workflows composed files whole, CWL GitHub repository. These snapshotted single ZIP, indicating workflow. [FDOF11] PID resolution content-negotiate Datacite’s [FDOF2] Signposting find containing [FDOF3] richer JSON-LD [FDOF5,FDOF8], see Fig. 1 Metadata schema.org [FDOF7] community-developed Bioschemas ComputationalWorkflow [FDOF10]. Workflows discovered API [FDOF5,FDOF6,FDOF11] created/modified CRUD [FDOF6] profile, effectively Type [FDOF7], declared https://w3id.org/workflowhub/workflow-ro-crate/1.0; language https://w3id.org/workflowhub/workflow-ro-crate#galaxy) Further work include formalise repositories [FDOF5,FDOF7], PIDs Signposting, HTTP navigation individual within shown adoption many disciplines, providing lightweight, therefore easy adopt, generating Objects. rapidly becoming integral part fabric between components demonstrated here WorkflowHub, contributing building European Open Science Cloud.