This step establishes a shared understanding of the types of research objects, components, domain concepts and relationships involved in the FAIRification activity, and selects or defines a domain model to guide the work.
A domain model describes the entities, concepts, properties, relationships and constraints that are relevant to a particular subject area or intended use. It provides a conceptual basis for later decisions about identifiers, metadata, standards, vocabularies, mappings and hosting. The model may already be explicit in a schema, metadata profile or community specification. It may instead be implicit in the structure, documentation or practices surrounding the existing research object. The project may reuse an established model, adapt or extend one, combine compatible models, or define a new model where no suitable option exists.
This step builds on the research object scope, component inventory, metadata, documentation, provenance and dependencies obtained in Step 1. It identifies what the current research object represents, what the intended uses require and where the current and required models differ. Adopting a domain model does not mean that the research object must be transformed immediately. This step makes and records the modelling decisions that will guide implementation in later Template steps.
Use this step during project examination to identify relevant types, models and requirements and to compare the current and required states. During an implementation cycle, use it to select, adapt or define the domain model that will guide the agreed FAIRification work.
Identify research object types
Select or define a domain model
Identifier minting
Identifier discovery and reuse
Vocabulary discovery and selection
Vocabulary extension and development
Semantic annotation
Vocabulary management
Identify research object types
Identify and describe the types of research objects and components included in the FAIRification activity.
Research object type identification informs the selection of appropriate domain models, metadata profiles, identifier schemes, standards, vocabularies and target hosting environments. It also helps determine which relationships and dependencies must be preserved.
Distinguish between:
- Research object types, such as datasets, software, workflows, models, notebooks, protocols, and compound research objects.
- Components, such as files, records, modules, workflow steps, metadata documents and referenced resources.
- Domain entity types, such as samples, organisms, participants, observations, assays, images, sequences and variables.
- Representation types, such as schemas, file formats and serialisations, which are examined and implemented in later steps.
A research object may have more than one type or contain several different types of components. Classification should therefore describe the composition of the research object set rather than force each object into a single category.
Resources
4 ELIXIR Stories items
- Plant Sciences Community Showcase Cascade mapping from FT 2–6: integration of multiple resources, data harmonisation
- FAIRtracks: FAIRtracks and Omnipy - FAIRtracks interoperability story Model the domain: Learnings published on fairtracks.net and included as ELIXIR RIR
- Single Cell Omics Community Interoperability Showcase Example of a domain with few standards and little convergence
- Metabolomics Community & Interoperability Platform - On which subjects could they collaborate? Model the domain: Learnings published on fairtracks.net and included as ELIXIR RIR
4 FAIR Metroline items
- Analyse data semantics Clarifying the meaning of the data helps determine which data types are present and relevant.
- Create or reuse a semantic (meta)data model Defining concepts and structure provides a basis for identifying which data types need to be represented.
- Register structural metadata Describing the structure of a resource helps recognise and distinguish the data types involved.
- Assess availability of your metadata Showing which metadata is already available helps identify how data types are currently described.
2 FAIR Cookbook items
Select or define a domain model
Select, adapt or define a model that adequately represents the concepts, entities, relationships and constraints required by the FAIRification activity.
This connects the research object and domain model to practical identification requirements. It determines whether identifiers are needed for the research object set, individual objects, versions, components, domain entities and related resources, and whether existing identifiers can meet those needs.
The capability may be provided through expertise within the project or through support from data stewards, repository specialists, domain communities, identifier-service providers or other infrastructure operators.