Bundled Knowledge Graphs

ARK CLI ships with three pre-loaded biomedical knowledge graphs. Each graph provides a different perspective on biomedical knowledge and is automatically discovered from the data/ directory at startup.

PrimeKG

PrimeKG is a precision medicine knowledge graph that integrates 20+ biomedical resources into a unified graph. It contains detailed information about diseases, drugs, genes/proteins, biological processes, pathways, anatomy, and more.

PrimeKG covers a wide range of node types (diseases, drugs/treatments, genes/proteins, biological processes, molecular functions, cellular components, pathways, anatomical structures, and exposures) connected by relationship types such as drug indications, contraindications, gene-disease associations, protein-protein interactions, and pathway membership.

Best for: Drug repurposing exploration, gene-disease associations, multi-hop biomedical reasoning across diverse entity types.

AfriMedKG

AfriMedKG is a knowledge graph derived from the AfriMed-QA pan-African medical question-answering benchmark. It focuses on medical knowledge relevant to African populations and healthcare contexts, covering diseases, treatments, and medical concepts specific to or prevalent in African regions.

Best for: African-focused medical knowledge, healthcare question answering, region-specific biomedical exploration.

OptimusKG

OptimusKG is a modern multimodal knowledge graph for precision medicine. It represents the latest approach to biomedical KG construction with built-in data quality checks, reproducible pipelines, and support for multiple data modalities.

Best for: Modern biomedical KG exploration, multimodal data integration, reproducible knowledge graph construction.

Comparing the Graphs

Next Steps