Bio2RDF (Linked Biomedical RDF Network)
Overview
Bio2RDF is an open-source, large-scale semantic web integration project that provides the largest network of Linked Data for the life sciences. The project addresses a critical challenge in biomedical research: life scientists need to search for and integrate biological data from a multitude of independent databases with non-standard web and database interfaces and heterogeneous data formats.
Bio2RDF creates a network of coherent linked data by converting over 35 biomedical and life sciences databases into standardized RDF (Resource Description Framework) format, making them accessible through SPARQL endpoints and REST APIs. The project is maintained by Maastricht University’s Institute of Data Science and continues to serve as critical infrastructure for biomedical research and knowledge discovery.
Data Content and Scale
Database Coverage
Bio2RDF Release 3 (July 2014) encompasses approximately 11 billion RDF triples across 35+ integrated datasets, representing one of the largest networks of linked data for life sciences.
Core Integrated Databases (Full Content):
- MGI (Mouse Genome Informatics) - Mouse genetic data and annotations
- HGNC (Human Gene Nomenclature Committee) - Human gene names and symbols
- KEGG (Kyoto Encyclopedia of Genes and Genomes) - Pathways and biological processes
- Entrez Gene/NCBI Gene - Comprehensive gene information
- OMIM (Online Mendelian Inheritance in Man) - Genetic disorders and inheritance
- Gene Ontology (GO) - Biological process, cellular component, molecular function annotations
- OBO (Open Biomedical Ontologies) - 39+ integrated biomedical ontologies
- PDB (Protein Data Bank) - 3D protein structures
- ChEBI (Chemical Entities of Biological Interest) - Chemical compound data
On-Demand Linked Sources:
- UniProt (protein sequences and annotations)
- Reactome (biochemical pathways)
- Prosite (protein domains and patterns)
- PubMed (biomedical literature abstracts)
- GenBank (nucleotide sequences)
- PubChem (chemical compounds)
Release 3 Additions (Expanded Coverage):
- ClinicalTrials.gov (clinical trial information)
- dbSNP (single nucleotide polymorphisms)
- GenAge (aging-related genes)
- GenDR (dietary restriction genes)
- LSR (Life Span Database)
- OrphaNet (orphan diseases)
- SIDER (drug side effects)
- WormBase (C. elegans genomics)
- And 27+ additional linked sources via PharmGKB
Data Types Covered
Bio2RDF integrates diverse biomedical data types:
- Genes and Proteins: Gene sequences, protein structures, functional annotations
- Chemical Compounds: Drug information, chemical properties, molecular structures
- Diseases: Disease associations, genetic origins, clinical manifestations
- Pathways: Metabolic and signaling pathways, biological processes
- Literature: PubMed abstracts, citations, publication metadata
- Genomics: SNPs, mutations, genomic variations
- Protein Structures: 3D coordinates, structural classifications
- Protein-Protein Interactions: Interaction networks and relationships
- Gene-Disease Associations: Links between genes and clinical phenotypes
- Drug-Target Interactions: Pharmacological relationships
Data Access Methods
SPARQL Endpoint
Main SPARQL Endpoint: https://bio2rdf.org/sparql
- Powered by OpenLink Virtuoso 7.2.0 (version 07.20.3236)
- Full W3C SPARQL 1.1 protocol support
- CORS-enabled for cross-origin requests
- Text-indexed for full-text search capabilities
- Supports federated queries using SPARQL 1.1 SERVICE keyword
Query Example:
curl -H "Accept: text/plain" --data-urlencode query@query.sparql https://bio2rdf.org/sparql
REST API
REST API Endpoint: Available through dedicated repository
- RESTful interface to Bio2RDF network of data
- Database available at http://dataset.bio2rdf.org/
- Beta version at http://beta.dataset.bio2rdf.org/
- GitHub repository: https://github.com/bio2rdf/bio2rdf-api
- Support for batch operations and on-demand queries
Data Downloads
Download Service: https://download.bio2rdf.org/
- Complete RDF files in multiple serialization formats
- Compressed files for efficient transfer
- Text-indexed Virtuoso triple stores for download and local deployment
- Version-controlled releases with metadata
Bio2RDF supports multiple output formats for flexibility in data consumption:
- SPARQL_Results_CSV - CSV format for spreadsheet applications
- SPARQL_Results_JSON - JSON for programmatic access
- N-Triples - Line-based RDF format
- Turtle - Human-readable RDF syntax
- N3 (Notation 3) - Extended RDF notation
- RDF/XML - XML serialization of RDF
Data Standards and Technical Architecture
Semantic Web Technologies
RDF (Resource Description Framework)
- Triple-based knowledge representation (Subject-Predicate-Object)
- URI-based entity identification using International Resource Identifiers (IRI)
- Standard format:
http://bio2rdf.org/namespace:id
- Normalized URIs for consistent entity linking across datasets
OWL (Web Ontology Language)
- Dataset types and predicates declared as OWL classes and object properties
- Enables semantic reasoning over implicit knowledge
- Support for subclass relations (rdfs:subClassOf), property equivalence (owl:equivalentProperty)
- Consistency checking and data quality validation
SPARQL (SPARQL Protocol and RDF Query Language)
- W3C standard for querying RDF data
- Complex graph pattern matching across linked datasets
- Federated query support across multiple endpoints
- Full-text search integration
Semanticscience Integrated Ontology (SIO)
Bio2RDF uses the Semanticscience Integrated Ontology (SIO) as its primary semantic framework:
- Upper-level ontology providing consistent types and relations
- Provides vocabulary for linked data projects across biological domains
- Covers physical entities, processes, and informational entities
- Includes extensions for chemistry, biology, biochemistry, and bioinformatics
- Freely available under Creative Commons Attribution license
Data Integration Pipeline
Bio2RDF’s ETL (Extract-Transform-Load) process:
- Data Extraction: Download from source databases or APIs
- Format Conversion: Transform from diverse formats (CSV, TSV, RDB, XML, JSON) to RDF
- URI Normalization: Generate consistent IRIs using centralized dataset registry
- Semantic Mapping: Map to SIO ontology and OBO ontologies
- RDF Generation: Produce RDF in multiple serialization formats
- Triplestore Loading: Index and store in OpenLink Virtuoso
- Quality Assurance: Consistency checking and validation
Transformation Framework: Modern approach using Common Workflow Language (CWL) for reproducible, scalable data transformation workflows
Technical Infrastructure
Core Technologies
- Triple Store: OpenLink Virtuoso Open Source Community Edition
- Scripting Languages: PHP, Perl, Ruby
- Containerization: Docker for deployment and scaling
- Version Control: GitHub for code and configuration management
- Text Indexing: Full-text search capabilities in Virtuoso
Architecture Components
Data Conversion Pipeline
- Central GitHub repository consolidating ~30 PHP scripts, 1 Java program, 1 Ruby gem
- Common API for IRI generation using centralized dataset registry
- Community contribution model with code review
SPARQL Endpoint Infrastructure
- Each dataset has its own SPARQL endpoint powered by Virtuoso
- Federated query support for cross-dataset queries
- Text-indexed for efficient full-text search
API and Interface Layer
- REST API for programmatic access
- Web-based browsing and discovery interface
- OpenLink Virtuoso SPARQL Query Editor for interactive queries
- BioSearch semantic search engine for user-friendly discovery
Use Cases and Applications
Drug Discovery and Repurposing
- Drug Repurposing: Identify existing drugs with new therapeutic applications
- Target Identification: Find novel therapeutic targets for diseases
- Polypharmacology Analysis: Understand multi-target drug effects
- Adverse Reaction Mapping: Identify potential drug side effects through cross-database analysis
Disease Research and Mechanistic Understanding
- Disease-Gene-Protein Networks: Map relationships between genes, proteins, and diseases
- Pathway Analysis: Understand biological mechanisms underlying diseases
- Comorbidity Discovery: Identify disease associations and relationships
- Genetic Disorder Understanding: Link molecular mechanisms to clinical phenotypes
Complex Biological Queries
Example Use Cases:
- Identifying chemicals participating in specific biological processes with protein-encoding interactions
- Exploring implications of transcription factor genes in disease pathways
- Determining protein interaction networks from time-course genomic data
- Mining relational paths in integrated biomedical data for knowledge discovery
Chemical-Biological Integration
- Chemical Biology: Integration of small molecule, target, gene, and pathway information (via Chem2Bio2RDF)
- Structure-Activity Relationships: Link chemical properties to biological activities
- Chemical Genetics: Connect chemical perturbations to genetic effects
Research Support
- Hypothesis Generation: Generate testable hypotheses through integrated data mining
- Data Mashups: Create custom knowledge graphs from Bio2RDF data
- Knowledge Discovery: Novel association discovery through linked data exploration
- Reproducible Research: Standardized RDF representation supports reproducibility
Organizational Structure
Leadership and Affiliations
Michel Dumontier
- Scientific Director and Distinguished Professor of Data Science
- Maastricht University Institute of Data Science
- Primary developer and maintainer of Bio2RDF
- Expertise in semantic web, biomedical ontologies, data integration
François Belleau
- Founder of the Bio2RDF project
- Original architect of the linked data framework
Community Contributors
- Alison Callahan, José Cruz-Toledo, Philippe Rigault (Bio2RDF Release 2)
- Vincent Emonet (Recent contributions)
- Active community of developers and researchers
Host Institution
Maastricht University Institute of Data Science (IDS)
- Primary institutional base for Bio2RDF maintenance and development
- Provides infrastructure and funding support
- Leads ongoing improvements and dataset updates
- GitHub organization: MaastrichtU-IDS
- Open-source, community-driven project
- Distributed development model with contribution mechanisms
- Code review and quality assurance processes
- Welcome to external contributions and dataset additions
Funding and Support
Primary Support: Maastricht University Institute of Data Science
Development Model: Supported through:
- University research funding
- Open-source community contributions
- Academic partnerships
- Grant-supported research projects
Status: Actively maintained with ongoing updates to datasets, infrastructure improvements, and feature enhancements
Standards Compliance and Interoperability
- W3C Standards: RDF, OWL, SPARQL, HTTP/REST
- Semantic Web Technologies: Linked data principles, URI-based identification
- Biomedical Standards: Integration with OBO Foundry ontologies
- Open Science: CC-BY 3.0 licensed, freely accessible data and code
- Interoperability: Support for federated queries and data integration
Citation and Usage
Bio2RDF data and services are freely accessible through public endpoints with no restrictions. The project is covered under the Creative Commons Attribution 3.0 License.
Recommended Citation
For Bio2RDF:
“Bio2RDF. Available at: https://bio2rdf.org”
For specific publications:
“Callahan A, Cruz-Toledo J, Ansell P, Dumontier M. Bio2RDF Release 2: Improved Coverage, Interoperability and Provenance of Life Science Linked Data. Journal of Biomedical Semantics. 2013;4(Suppl 1):S1.”
Additional Resources
- Main Website: https://bio2rdf.org/
- SPARQL Endpoint: https://bio2rdf.org/sparql
- REST API: https://github.com/bio2rdf/bio2rdf-api
- Downloads: https://download.bio2rdf.org/
- GitHub Repository: https://github.com/MaastrichtU-IDS/bio2rdf
- Scripts Repository: https://github.com/bio2rdf/bio2rdf-scripts
- BioSearch Search Engine: https://biosemantics.org/biosearch
- SIO Ontology: https://bioportal.bioontology.org/ontologies/SIO
- Maastricht University IDS: https://www.maastrichtuniversity.nl/research/institutes/institute-data-science
Bio2RDF continues to serve as critical infrastructure for integrating biomedical knowledge, enabling researchers worldwide to discover novel relationships, generate hypotheses, and advance biomedical science through standardized semantic web technologies.