Skip to content

FAIR² Method Representation

This document defines how FAIR² expresses detailed methods using structured, machine-readable JSON-LD constructs. It is fully compatible with schema.org, ML Croissant, and the FAIR² metadata specification.

Overview

Method descriptions in FAIR² are structured hierarchically using the following types:

  • fair2:Section derived from schema:HowToSection
  • fair2:Step derived from schema:HowToStep
  • fair2:Substep derived from schema:HowToStep
  • fair2:StepCase derived from schema:HowToStep

Each level extends or specializes the schema.org/HowToStep pattern, enabling context-rich and modular method representation.

Structure and Semantics

Section (fair2:Section = schema:HowToSection)

A major methodological unit (e.g., "Structure Prediction", "Binding Assays").

{
  "@type": "HowToSection",
  "@id": "#section-structure-prediction",
  "name": "Structure Prediction",
  "steps": [ ... ]
}

Step (fair2:Step = schema:HowToStep)

Primary procedural element. Must contain text, and can include substeps or conditions.

{
  "@type": "HowToStep",
  "@id": "#step-alphafold",
  "text": "Run AlphaFold prediction using ColabFold pipeline.",
  "substeps": [ ... ],
  "conditions": [ ... ]
}

Substep (fair2:Substep = schema:HowToStep)

A finer-grained instruction within a Step.

{
  "@type": "HowToStep",
  "@id": "#substep-prepare-input",
  "text": "Prepare FASTA input and validate sequence."
}

StepCase (fair2:StepCase = schema:HowToStep)

Used to capture conditional logic (e.g., environment-dependent, tool-specific alternatives).

{
  "@type": "HowToStep",
  "@id": "#stepcase-gpu",
  "name": "If GPU is available",
  "text": "Run full-db AlphaFold with GPU acceleration."
}

Ontological Alignment

Each element inherits from schema:HowToStep, but is semantically specialized via fair2: namespace identifiers. This ensures compatibility with:

  • schema.org-based search indexing
  • Croissant Method constructs
  • SHACL validation using RDF types

Integration with Activities

Steps may link to provenance-aware prov:Activity elements for detailed tracking:

"wasGeneratedBy": {
  "@type": "prov:Activity",
  "type": "Prediction",
  "wasAssociatedWith": { "@type": "prov:Agent", "name": "AlphaFold v2.3" }
}

Example JSON-LD Block

{
  "@type": "fair2:Section",
  "name": "Structure Prediction",
  "steps": [
    {
      "@type": "fair2:Step",
      "text": "Run AlphaFold prediction.",
      "substeps": [
        { "@type": "fair2:Substep", "text": "Prepare FASTA sequence input." },
        { "@type": "fair2:Substep", "text": "Launch ColabFold and upload sequence." }
      ],
      "conditions": [
        { "@type": "fair2:StepCase", "name": "GPU available", "text": "Use full-db AlphaFold." },
        { "@type": "fair2:StepCase", "name": "CPU only", "text": "Use reduced-db fallback mode." }
      ]
    }
  ]
}

SHACL Compliance

All method representations must conform to the MethodShape defined in FAIR² SHACL:

  • @type must be schema:HowToSection or schema:HowToStep
  • steps must be an array of schema:HowToStep
  • text is mandatory for all procedural steps

Notes

  • This pattern allows nesting to arbitrary depth.
  • Textual content can be multilingual using @language tags.
  • Conditions and substeps are optional but recommended for clarity and modularity.

This structured approach enables both human-readable and machine-actionable protocols across diverse scientific domains. For more examples, see the fair2.json specification.