FAIR-by-Design approach

F1. Globally unique, persistent identifiers

• HTTP URI per entity (e.g., https://api.ontaya.com/v2/Experiment/{id}) with content negotiation.
• Relay-style id: ID! as persistent key; DOIs/ARKs for published datasets.
  F2. Rich metadata descriptions
• Keywords, abstract, methods + JSON-LD contexts for harvesting.
  F3. Metadata includes data identifiers
• Embed @id in JSON-LD and return id prominently in GraphQL payloads.
  F4. Indexed, searchable resources
• Elastic/Solr index, faceted search, /search REST alongside GraphQL.

A1. Retrievability via standard protocol

• HTTPS REST + GraphQL; JSON and JSON-LD via Accept negotiation.
  A1.1. Open, free protocols
• HTTP(S)/GraphQL, documented via OpenAPI/introspection.
  A1.2. Auth where necessary
• OAuth2/JWT scopes (Keycloak/Auth0), Symfony security; public metadata when possible.
  A2. Metadata persistence
• Retain metadata even if raw data is archived; expose status/provenance.

I1. Formal, shared language

• JSON-LD context/RDF schema mapping GraphQL types to Schema.org/OBO; SPARQL for federation.
  I2. Community vocabularies
• NCBI Taxonomy, EDAM, OBO terms; validate on ingestion.
  I3. Qualified references
• Resolvable URIs with relation metadata; GraphQL edges with cursor + node.id; relation types in JSON-LD.

R1. Rich description

• Provenance (createdBy/At, modifiedBy/At), license, usage notes; domain fields (dose, protocol, environment).
  R1.1. License
• Top-level license (CC0/CC-BY) exposed machine-readably; UI license picker.
  R1.2. Detailed provenance
• Auto-track “who did what when” via triggers/events; expose provenance subtype in GraphQL.
  R1.3. Community standards
• Validate against ARRIVE/MIAME/other schemas at ingestion; UI templates enforce required fields.