biolink

Biolink is a community-driven data model and interoperability layer for biomedicine that aims to unify heterogeneous biological data into comparable graphs. It provides a standardized vocabulary of entity types and relationship types that enables researchers to integrate data from diverse databases and ontologies into a single knowledge graph.

The Biolink Model was developed by the Monarch Initiative and has since been adopted by multiple projects to support cross-database queries and comparative biology.

Core concepts include semantic types such as Gene, Protein, Disease, Phenotype, Chemical, Pathway, and AnatomicalEntity, as well as a hierarchy of predicates like related_to, has_phenotype, part_of, located_in, regulates, and interacts_with. The model is designed to cover common biological relationships while remaining extensible for new domains and data sources.

Identifiers and alignment: Biolink uses CURIE-style prefixes and maps to identifiers across major resources (HGNC, Ensembl, UniProt, MONDO, DOID, HPO, ChEBI, GO) and aligns with ontology and FAIR data practices. This supports consistent referencing and cross-resource integration.

Format, tools, and usage: The model is published with a formal specification and reference implementations; it supports RDF, OWL, and JSON-LD representations and is accompanied by libraries to validate and serialize data. It is maintained in public repositories with versioned releases and community input.

Impact: Biolink provides a common semantic substrate for biomedical knowledge graphs, enabling cross-species and cross-domain queries, data integration, and reproducible research. It underpins efforts to harmonize biological data and facilitate interoperable analyses across diverse data sources.