FAIR Principles

---

FAIR (Findable, Accessible, Interoperable, Reusable) is a set of principles originally written for research data (see Wilkinson et al 2016) but since expanded to other research output (see Baker et al 2022, Patel et al 2023). Both European and Swedish funders and universities currently recommend or require adhering to FAIR principles (e.g., Swedish Research Council: Making research data accessible and FAIR criteria). There is no one specific way to 'make something FAIR'; instead, research output can adhere to FAIR principles to different extent and in different ways.

SciLifeLab Serve helps machine learning models and data science applications that have been developed during research projects to be made available according to FAIR principles. Below we describe how exactly models and applications made available through our service adhere to FAIR principles. This page is heavily inspired by (and parts of the text are copied from) a similar page on Zenodo.

How SciLifeLab Serve helps to adhere to FAIR Principles

To be Findable:

• F1: (meta)data are assigned a globally unique and persistent identifier
  • A DOI is issued to every public application made available through SciLifeLab Serve.
• F2: data are described with rich metadata (defined by R1 below)
  • SciLifeLab Serve's metadata is compliant with DataCite's Metadata Schema mandatory and recommended terms. In a few cases we made additional metadata mandatory for DOIs issued through us, and in a few cases we additionally restricted the vocabularies to ensure standardisation.
• F3: metadata clearly and explicitly include the identifier of the data it describes
  • The DOI is a top-level and a mandatory field that is provided in the metadata of each application on SciLifeLab Serve.
• F4: (meta)data are registered or indexed in a searchable resource
  • Metadata of each public application is sent to DataCite servers during DOI registration and indexed there.
  • Selection of metadata of each public application is searchable directly in SciLifeLab Serve's app catalogue immediately after publishing.

To be Accessible:

• A1: (meta)data are retrievable by their identifier using a standardized communications protocol
  • Metadata for individual apps are harvestable using the DataCite REST API service (see all entries on DataCite Commons) by the DOI record identifier.
• A1.1: the protocol is open, free, and universally implementable
  • See point A1. REST is an open, free and universal protocol for information retrieval on the web.
• A1.2: the protocol allows for an authentication and authorization procedure, where necessary
  • Metadata are publicly accessible and licensed under public domain. No authorization is ever necessary to retrieve it.
• A2: metadata are accessible, even when the data are no longer available
  • The metadata will be retained for the lifetime of the SciLifeLab Serve. This is currently at least the duration of the SciLifeLab & Wallenberg National Program for Data-Driven Life Science, which is funded until 2032.
  • The metadata should be available even after that point in time through DataCite.

To be Interoperable:

• I1: (meta)data use a formal, accessible, shared, and broadly applicable language for knowledge representation.
  • DataCite offers retrieval of metadata in DataCite XML, DataCite JSON, Schema.org JSON-LD formats.
• I2: (meta)data use vocabularies that follow FAIR principles
  • SciLifeLab Serve uses controlled vocabularies for the subjects and keywords (specifically: MeSH Medical Subject Headings, EuroSciVoc The European Science Vocabulary, and GEMET General Multilingual Environmental Thesaurus), creator identifiers (ORCID), creator affiliation (ROR), funder identifiers (ROR).
• I3: (meta)data include qualified references to other (meta)data
  • Each referenced external piece of metadata is qualified by a resolvable URL.

To be Reusable:

• R1: (meta)data are richly described with a plurality of accurate and relevant attributes
  • Each app metadata record contains a minimum of DataCite's Metadata Schema mandatory terms, with optionally additional DataCite's Metadata Schema recommended terms and SciLifeLab Serve's own enrichments.
• R1.1: (meta)data are released with a clear and accessible data usage license
  • [We are currently working on this.]
• R1.2: (meta)data are associated with detailed provenance
  • All metadata uploaded is traceable to a registered SciLifeLab Serve user.
• R1.3: (meta)data meet domain-relevant community standards
  • SciLifeLab Serve is not a domain-specific repository. Through compliance with the DataCite's Metadata Schema, metadata that we collect and store meet one of the broadest cross-domain standards available.