Ontograph

About

Ontograph is an HPC-powered pipeline designed to integrate and unify biomedical knowledge graphs from a multitude of EMBL-EBI, MONARCH Initiative, ROBOKOP, and other key resources. Its primary goal is to simplify complex, integrative queries for both LLM agents and human users, overcoming the limitations of single-resource REST APIs. By employing ontologies and LLM embeddings, it constructs a cohesive knowledge base, providing outputs as both periodically materialised query results (CSV) and downloadable Neo4j database exports, accessible via a development server and MCP endpoint.

Key Features

• Provides pre-built Neo4j database exports for powerful querying
• HPC pipeline for aggregating diverse biomedical knowledge graphs
• 4 GitHub stars
• Offers a Streamable HTTP MCP endpoint for programmatic access
• LLM embeddings and ontology-driven data integration strategies
• Materialises graph queries into downloadable CSV tables

Use Cases

• Performing integrative queries across multiple biomedical resources using LLM agents or human analysts
• Analyzing large-scale biomedical data by loading materialised query results into data processing libraries
• Setting up local or HPC-based Neo4j instances for custom, high-performance graph querying