Each registry within FAIRsharing has a number of types.
Standards records are divided into five subtypes. In order to be approved by the FAIRsharing team, the standards must be created by or have a strong backing from a community. This is commonly a grassroots or community-led standard, but may also be from a formal standards development organisation (SDO).
Please choose the standard type appropriate for your resource. If your resource does not fit into any of these categories, then please get in touch:
- Model and Format records describe resources that provide a common structure for all data of a given experimental type. They are structured representations of information from a conceptual model or schema, which enables (syntactic) interoperability among systems. These formats are often created using base formats such as XML, CSV or RDF.
- Terminology artefacts are controlled vocabularies or ontologies that describe the meaning, or semantics, of data and metadata. They provide a method of annotating data in a way that is useful for humans and comprehensible by computers, and allow for better collation and searching of the data.
- Reporting guidelines provide a checklist for the minimal descriptive content required to evaluate, interpret and disseminate an experiment. Defining a common set of metadata to guide researchers in reporting scientific context is important for data repositories, journals and funders. Often reporting guidelines do not have a literal checklist, but are rather presented as a document stating required and optional attributes.
- Identifier schemas (e.g. URLs, pURLs, DOIs) describe publicly-available identifiers from a central registry. For more information on this type, see the subsection below.
- Metrics are criteria to assess some aspect of a digital resource, for example - but not limited to - its discoverability. As an example, the FAIR Metrics (of which FM-I2 is one) are a set of machine-actionable criteria that allow the assessment of the level of FAIRness of a resource.
Identifier schema records are intended to describe those publicly-available, persistent, and unique identifiers used and shared across multiple resources. Identifiers that are only ever intended to be minted for a single repository, e.g. UniProt IDs. The examples below give some ideas about what is and is not accepted:
- 1.ID schemas in scope for this record type. DOI, Handle, and ARK are schemata for resolvable PIDs that are used in cross-community, multi-disciplinary ways, and are used and assigned across multiple projects. Please note that some of our records seem to be, at first glance, for a single resource (IVOA identifiers). However, they are in scope because IVOA is a large consortium of virtual observatories that all use this identifier system for linking and identifying data.
- 2.ID schemas in scope for this record type. Non-resolvable, but persistent and unique identifiers shared across resources and communities such as Reaction InChi, CAS registration numbers and EC numbers. Although they don't "click and resolve" without help, they are consistently used through a number of communities and therefore are in scope.
- 3.ID schemas NOT in scope for this record type. Identifiers only ever minted for a single repository, e.g. UniProt IDs. Although often shared as cross-references and links back to the originating resource, they are not intended to be created for anything other than that resource's contents.
Database records are divided into three subtypes. Please choose the record type appropriate to your resource. If your resource does not fit into any of these categories, then please get in touch. Please choose from the following types:
- Repositories allow the submission, storage of and access to data (datasets, software, materials and other digital objects) and include (multi)disciplinary, generalist, project, data type and institutional repositories
- Knowledgebases synthesise data from a number of sources including published literature, databases and other types of data sources. They often have a manual curation component. They can be considered secondary databases as they store data derived from primary sources.
- Knowledgebases and Repositories are databases that have features of both knowledgebases and repositories.
While "Database" is used as the name of this registry, in this regard please note that we have a very generous definition of "data". Anything that stores digital objects (e.g. software repositories, digital representations of physical collections) and that meets the conditions outlined below, may be described in FAIRsharing.
It is important to note that FAIRsharing database records are intended to describe resources that store datasets; we do not accept submissions describing the datasets themselves. FAIRsharing is a registry of linked information on content standards, databases, and data policies in a variety of research areas, and as such does not describe data sets or allow data upload. More information as to what constitutes a suitable resource for incorporation into FAIRsharing can be found on our website.
At FAIRsharing, one of our roles is to help you find the right database to upload your data. Try browsing our complete subject hierarchy or searching our database records to find an appropriate database for your data type. Alternatively, you could look at our collection of Generalist Repositories or browse all subject-agnostic records, which may provide you with a location to store data from a wider range of research areas. If you are submitting to a particular journal, you can also search for that journal or its publisher; if they are registered with us, then their FAIRsharing data policy record may contain a list of recommended databases and/or standards.
Here are the criteria for acceptance as a FAIRsharing database record:
- The resource should be an organised collection of data (see our definition of "data" earlier in this section) and datasets rather than just an individual dataset.
- The resource is findable. Users can access the database via an active website and can also browse and/or search the database. In contrast, datasets are generally downloadable but not searchable, and therefore are not appropriate for FAIRsharing.
- The resource is accessible. Irrespective of licence type, the resource is available to users via a dedicated website (even if a log in or payment is required).
FAIRsharing stores descriptions of data policies from a number of sources outlined below. The common feature of all of them is that the data policy should describe how each source recommends the storage and/or formatting of data.
- Journal: A journal data policy describes those formats and/or databases that are recommended for use for authors submitting a manuscript to that journal. Such resources may be described either explicitly or in a more general way.
- Society: Particular societies (e.g. standards development organisations) often provide data policies so that researchers wishing to align with society best practices have a concrete data policy they can implement.
- Funder: Funders often have strict requirements about the type of databases or standards that they wish their grantees to use.
- Project: Particular projects may have a data policy that describes those resources that must be used by project members. Please note that this policy type should only be used for large-scale projects.
- Institution: This is a relatively new policy type for FAIRsharing. Please use this to describe your institution's (e.g. a university's) data policy for its employees and, sometimes, collaborators.
- Journal Publisher: Sometimes the publisher of a stable of journals has overarching data policies that apply to all journals.
Irrespective of the type of policy you are describing, it is important that you provide accurate and complete metadata to describe the policy itself. Of particular interest are the relationships you can add to your policy record with us; these provide direct links from your policy record to the database and standards records that your policy recommends. Also, data policies can be nested, allowing one data policy to extend another data policy(ies). This is useful in many situations, such as when a journal publisher has an overall data policy, and then individual journals from that publisher have their own specific extension of that policy.
Further information below shows how FAIRsharing is working to align with a number of different policy metadata standardisation efforts. This means that registering your policy with us provides prospective users with human- and machine-readable metadata describing the characteristics of your policy.
Collections are containers that allow you to create branded "slices" of FAIRsharing records that are specific to your needs. The infographic below provides a few examples of the ways in which our user community highlights their resources of interest using our collections. The examples in the image are IVOA, the RDA COVID-19 WG Resources, CDISC, and a crosswalk of metadata schemas and guidelines.
A few examples of the ways collections can help your organisation or community.