CAM-KP (Causal Activity Models Knowledge Provider)
Overview
CAM-KP (Causal Activity Models Knowledge Provider) is a web service knowledge graph that provides access to integrated causal biological and chemical models. As part of the NCATS Biomedical Data Translator initiative, CAM-KP serves as a critical infrastructure component for connecting siloed biomedical datasets through standardized semantic representations.
The knowledge graph integrates multiple types of causal models—including manually curated Gene Ontology Causal Activity Models (GO-CAMs), automatically derived Reactome pathway models, and chemical-gene interactions from the Comparative Toxicogenomics Database (CTD)—into a unified, OWL-based semantic platform accessible via both SPARQL and TRAPI-compliant REST APIs.
Data Content and Scale
Model Types and Coverage
CAM-KP combines three primary sources of causal models:
Gene Ontology Causal Activity Models (GO-CAMs)
- Manually annotated by Gene Ontology expert biocurators
- Over 750,000 experimental gene annotations from 150,000+ distinct scientific publications
- Link genes, gene products, and biological processes in causal networks
- Structured using Gene Ontology controlled vocabularies
Reactome Pathway CAMs
- Complete set of normal human Reactome pathways converted to GO-CAM representation
- Reactions translated to gene product activities
- Causal relations inferred between molecular activities
- Enables pathway reasoning through GO-CAM framework
Comparative Toxicogenomics Database (CTD) Interactions
- 3.8+ million direct chemical-gene interactions
- 17,700+ chemicals covered
- 55,400 genes represented
- 7,200 diseases linked to chemical exposures
- 214,000+ documented exposures
Knowledge Representation
- RDF Triplestore: Triple format knowledge graph with semantic richness
- OWL 2 Ontology Layer: Enables formal semantic reasoning over implicit knowledge
- Merged Bio-Ontology: Integrated vocabulary from 39+ OBO Library ontologies
- Precomputed Inferences: OWL-based reasoning results stored for efficient query response
Query Interfaces
SPARQL Endpoint
- Direct semantic queries: https://stars-app.renci.org/cam/sparql
- Full W3C SPARQL support
- Complex graph pattern matching over RDF triples
- Suitable for advanced semantic queries
TRAPI-Compliant REST API
- RESTful web service: https://cam-kp-api-dev.renci.org/1.2.0/query
- JSON request/response format
- Standardized Translator Reasoner API (TRAPI) compliance
- Enables federated querying across Translator ecosystem
Automat Integration
- Discovery and integration interface: https://automat.renci.org/#/cam-kp
- Part of broader RENCI knowledge provider integration platform
- Simplified access for researchers unfamiliar with APIs
Core Semantic Standards:
- RDF (Resource Description Framework) - Triple-based knowledge representation
- OWL 2 (Web Ontology Language) - Semantic expressiveness and reasoning
- SPARQL - Graph query language
- JSON - API response format with JSON-LD context
- TRAPI - Translator Reasoner API standard
Data Organization:
- Biolink Model - Universal schema for biomedical knowledge graphs with standardized entity types and relationships
- OBO Format - Integration with 39+ Open Biomedical Ontologies
- CURIE Representation - Compact URIs for cross-database entity identification
Data Sources and Integration
Primary Contributors
Gene Ontology Consortium
- Gene Ontology Causal Activity Models (GO-CAMs)
- Expert biocurator annotations
- Controlled vocabularies and structured annotations
Reactome Database
- Entire set of human biochemical pathway models
- Converted to GO-CAM representation via Pathways2GO tool
- Enables systems-level biological reasoning
Comparative Toxicogenomics Database (CTD)
- Manually curated chemical-gene interactions
- Chemical-disease associations
- Gene-disease relationships
- Cross-species data integration
OBO Foundry Ontologies
- Basic Formal Ontology (BFO) for foundational concepts
- Relation Ontology for relationship definitions
- Domain-specific ontologies for specialized knowledge
- Total of 39+ integrated OBO ontologies
Data Integration Pipeline
The cam-pipeline ETL system processes raw data through:
- Data extraction from source databases
- Standardization to Darwin Core terms where applicable
- RDF/OWL conversion
- Ontology alignment and semantic harmonization
- OWL 2 inference and reasoning
- Triplestore loading and indexing
Technical Architecture
System Components
Data Ingestion
- cam-pipeline: ETL framework for automated data loading
- Pathways2GO: Tool for converting Reactome pathways to GO-CAM format
- Bulk loaders for Gene Ontology and CTD data
Knowledge Representation Layer
- RDF triplestore backend
- OWL 2 ontology with precomputed reasoning
- Integrated vocabulary from 39+ OBO ontologies
- Support for semantic web standards
Query Infrastructure
- SPARQL endpoint for direct semantic queries
- TRAPI-compliant REST API for standardized access
- JSON-LD context management for CURIE expansion
- Support for federated query composition
Integration
- NCATS Translator ecosystem participation
- Biolink Model compliance
- OpenAPI/Swagger documentation
- Automat platform integration
Technologies Used
- Semantic Web: RDF, OWL, SPARQL
- Knowledge Representation: OBO ontologies, Gene Ontology, Biolink Model
- API Standards: TRAPI (Translator Reasoner API), OpenAPI
- Infrastructure: RENCI Stars platform hosting
- Version Control: GitHub repositories
Use Cases and Applications
Research Applications
Mechanistic Hypothesis Generation
- Propose molecular pathways connecting chemicals to disease
- Identify target pathways for drug discovery
- Support environmental health research with mechanistic understanding
Drug Discovery and Repurposing
- Identify potential therapeutic targets
- Discover new applications for existing drugs
- Generate biomedical hypotheses through integrated reasoning
Chemical Safety Assessment
- Understand toxicological mechanisms of action
- Predict adverse outcomes from chemical exposures
- Support regulatory decision-making
Translational Research
- Bridge molecular discoveries to clinical applications
- Generate testable hypotheses for experimental validation
- Support precision medicine research
Environmental Health Research
- Link environmental exposures to health outcomes
- Identify vulnerable populations and biomarkers
- Support exposure science studies
Systems Biology
- Model complex biological processes in causal networks
- Connect molecular, cellular, and organism-level effects
- Support hypothesis-driven research
Integration with Other Resources
CAM-KP operates as part of the NCATS Data Translator ecosystem, integrating with:
- ICEES (Integrated Clinical and Environmental Exposures Service) for clinical-environmental data
- Gene Ontology for curated biological annotations
- Reactome for pathway knowledge
- CTD for chemical-gene-disease associations
- Other NCATS Translator Knowledge Providers for federated queries
Organizational Structure
Leadership
Jim Balhoff (RENCI, UNC Chapel Hill)
- Assistant Director of Analytics & Data Science
- Expertise in bio-ontologies, semantic web, data integration, bioinformatics
- ORCID: 0000-0001-7695-6090
Stephen Edwards (RTI International)
- Director at US Environmental Protection Agency
- Expertise in data mining, knowledgebase design, ontology-based modeling
Host Institution
RENCI (Renaissance Computing Institute)
- Part of University of North Carolina at Chapel Hill
- Provides infrastructure, hosting, and technical development
- Participates in NCATS Biomedical Data Translator program
Collaborating Organizations
- Gene Ontology Consortium - GO-CAM curation and maintenance
- RTI International - Development and environmental expertise
- NCATS (National Center for Advancing Translational Sciences) - Funding and program oversight
Funding and Support
Primary Funder: National Center for Advancing Translational Sciences (NCATS), National Institutes of Health (NIH)
Program: NCATS Biomedical Data Translator Initiative
- Launched in 2016 as multiyear, iterative program
- Supports nationwide biomedical research data integration
- Consortium funding: approximately $13.5 million in FY 2020
- NCATS FY 2024 budget: $928,323,000 overall
Status: Repository archived May 23, 2024 (read-only); service continues to operate with active maintenance and support
Citation and Usage
CAM-KP data and services are freely accessible through public endpoints with no licensing restrictions. Data from integrated sources (Gene Ontology, Reactome, CTD) follow their respective open licenses (typically CC-BY or CC0).
Recommended Citation
For CAM-KP:
“CAM-KP (Causal Activity Models Knowledge Provider). Available at: https://automat.renci.org/#/cam-kp”
For the foundational work:
“Balhoff JP, Bizon C, Carlson J, et al. A Biomedical Knowledge Graph System to Propose Mechanistic Hypotheses for Real-World Environmental Health Observations. JMIR Medical Informatics. 2021;9(7):e26714.”
Additional Resources
- Main Access: https://automat.renci.org/#/cam-kp
- SPARQL Endpoint: https://stars-app.renci.org/cam/sparql
- REST API: https://cam-kp-api-dev.renci.org/1.2.0/query
- Documentation: https://github.com/ExposuresProvider/cam-kp-api/wiki
- GitHub Repository: https://github.com/ExposuresProvider/cam-kp-api
- Data Pipeline: https://github.com/ExposuresProvider/cam-pipeline
- Contact: Jim Balhoff (balhoff@renci.org)
CAM-KP continues to serve as a critical infrastructure component for the NCATS Biomedical Data Translator, enabling researchers to generate mechanistic hypotheses, support drug discovery efforts, and advance environmental health research through integrated access to curated biomedical knowledge.
Automated Evaluation