FAIR (guiding principles)

FAIR (guiding principles)
Name	FAIR (guiding principles)
Introduced	2016
Authors	Mark D. Wilkinson, Susanna-Assunta Sansone, Brian L. Doan
Purpose	Data stewardship
Related	Open science, Digital preservation, Research data management

FAIR (guiding principles) FAIR is a set of guiding principles for enhancing the reuse of digital assets, introduced to improve data stewardship across scientific, governmental, and commercial institutions. It complements movements such as Open access, Data citation, and Open Researcher and Contributor ID while interacting with policy frameworks from bodies like the European Commission, National Institutes of Health, and Organisation for Economic Co-operation and Development. The principles aim to make digital objects more discoverable, interoperable, and reusable across infrastructures exemplified by projects such as Dataverse, Zenodo, and Dryad.

Overview

The FAIR principles were first articulated by a group of stakeholders convened through initiatives associated with Force11, GO FAIR, and researchers including Mark D. Wilkinson and Susanna-Assunta Sansone. They were motivated by challenges encountered in large-scale efforts led by organizations such as European Bioinformatics Institute, National Aeronautics and Space Administration, and World Health Organization where datasets from initiatives like Human Genome Project, CERN, and UK Biobank required standardized stewardship. FAIR is technology-agnostic and complements standards from bodies including World Wide Web Consortium, International Organization for Standardization, and Research Data Alliance.

Principles (Findable, Accessible, Interoperable, Reusable)

Findable: Digital objects should be locatable using persistent identifiers such as Digital Object Identifiers and resolvers like Handle System. Metadata should reference authorities akin to Library of Congress, ORCID, or repositories exemplified by Figshare to support discovery in platforms like Google Scholar and PubMed.

Accessible: Access protocols should be standardized and documented, leveraging methods endorsed by Internet Engineering Task Force or infrastructures like OAuth and HTTP. Access models may involve stewardship policies from entities such as European Research Council or National Science Foundation while preserving data protection frameworks like General Data Protection Regulation.

Interoperable: Data and metadata should employ shared vocabularies, ontologies, and formats endorsed by communities including Gene Ontology Consortium, Dublin Core, and Simple Knowledge Organization System. Interoperability is advanced through standards from World Wide Web Consortium and semantic frameworks used by projects like Linked Data and Semantic Web.

Reusable: Metadata and licensing must enable reuse, often via licenses from Creative Commons or data use agreements used by ClinicalTrials.gov and repositories such as OpenAIRE. Provenance practices from initiatives like PROV (W3C) and citation norms promoted by DataCite increase trust and reproducibility in workflows like those in Large Hadron Collider analyses.

Implementation and Best Practices

Adoption of FAIR involves institutional policy, metadata standards, and technical infrastructures. Universities such as University of Oxford and consortia like European University Association integrate FAIR into research data management plans modeled on templates from Horizon 2020 and mandates from funders including Wellcome Trust and Bill & Melinda Gates Foundation. Best practices include minting persistent identifiers via DataCite, applying vocabularies from Global Biodiversity Information Facility, and using repository platforms like Zenodo or Figshare for deposition. Automation techniques using tools associated with GitHub, Jupyter Notebook, and Docker containers support reproducible packaging, while metadata registries maintained by FAIRsharing and indexing by aggregators like Crossref facilitate discovery.

Impact on Research Data Management

FAIR has influenced policy and infrastructure across sectors, shaping guidelines from European Commission programs, National Institutes of Health data policies, and national strategies in countries such as United Kingdom and United States. It has catalyzed interoperability efforts in domains from genomics (linked to European Bioinformatics Institute) to climate science (linked to World Meteorological Organization datasets), and spurred development of training curricula by organizations such as The Carpentries and ELIXIR. Implementation has shown benefits in enabling large-scale synthesis projects like those at Global Biodiversity Information Facility and collaborative experiments at facilities such as CERN.

Criticisms and Limitations

Critiques highlight ambiguities and implementation burdens: small institutions and projects may struggle with costs and skills compared with well-resourced actors like Wellcome Trust-funded centers. Concerns arise about overemphasis on machine-actionability versus human contexts noted by commentators from Nature and policy analysts from Organisation for Economic Co-operation and Development. Legal and ethical constraints under regimes such as General Data Protection Regulation and national privacy laws complicate openness. Additionally, differing community standards—seen across neuroscience consortia and astronomy archives such as European Southern Observatory—challenge uniform interpretation.

Governance, Standards, and Tools

Governance of FAIR-related activities is distributed across organizations: GO FAIR, Research Data Alliance, W3C, and national infrastructures like UK Research and Innovation coordinate standards and stewardship. Tools and registries that operationalize FAIR include DataCite, Crossref, FAIRsharing, BioSchemas, and ontologies hosted by OBO Foundry. Certification efforts such as CoreTrustSeal and frameworks from ISO bodies help assess repository readiness. Training and community engagement from The Carpentries, ELIXIR, and university libraries underpin capacity building for FAIR implementation.

Category:Data management