SPOKE: A massive biomedical knowledge graph for precision health and drug discovery

2026-02-19 · Source: Ai2 · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Advanced, extended

Summary

SPOKE (Scalable Precision Medicine Open Knowledge Engine) is a massive biomedical knowledge graph developed by Sergio Barzini and colleagues at UCSF, integrating data from over 65 human-curated specialized databases. It contains approximately 150 million semantically specified relationships, spanning entities from genes and proteins to pathways, molecules, exposures, and even food, organized across 11 ontologies. Built on a Neo4j property graph framework, SPOKE is continuously updated weekly. The project, funded in part by an NSF Convergence Accelerator program, aims to formalize and compute over biomedical relationships to address the complexity of biological data integration. Applications include drug repurposing, such as identifying bexarotene for multiple sclerosis myelination, and precision medicine, where patient electronic health record data is embedded into the graph to predict disease onset, achieving 83% accuracy for MS three years prior to diagnosis.

Key takeaway

For AI Scientists and Research Scientists working on biomedical applications, SPOKE demonstrates the power of integrating diverse, curated data into a knowledge graph. You should consider adopting a similar graph-based approach to formalize complex biological relationships, especially when tackling drug repurposing or early disease prediction. The ability to embed patient data and use LLMs for explainable predictions offers a robust framework for augmenting traditional research and clinical decision-making, potentially accelerating discoveries and improving patient outcomes.

Key insights

SPOKE integrates vast biomedical data into a knowledge graph for drug discovery and precision medicine.

Principles

• Data needs context to become information, and links to become knowledge.
• Integrate diverse biomedical data across scales for comprehensive patient assessment.
• Knowledge graphs enable formal computation over complex biological relationships.

Method

SPOKE uses a graph-theoretic approach to identify potential drug candidates by analyzing causal paths between compounds and biological processes, prioritizing drugs with higher probabilities of influencing target pathways.

In practice

• Embed patient EHR data into knowledge graphs for enhanced predictions.
• Use graph analysis to identify drug repurposing candidates.
• Leverage LLMs with RAG or MCP for conversational biomedical queries.

Topics

• Biomedical Knowledge Graph
• Drug Repurposing
• Precision Medicine
• Large Language Models
• Early Disease Prediction

Best for: AI Scientist, Research Scientist, AI Engineer, Data Scientist, AI Researcher

Related on AIssential

• Knowledge Graphs and Reasoning LLMs for Finding Simple Yet Effective Transcriptomic Perturbation Predictors — Simple K-nearest neighbour models leveraging knowledge graphs, enhanced by RL-trained LLMs, offer competitive transcriptomic perturbation prediction.
• ChatHealthAI: Aligning Electronic Health Record Representations with Large Language Models for Grounded Clinical Reasoning — ChatHealthAI integrates EHR representations with LLMs for grounded, interpretable clinical reasoning and accurate patient prediction.
• Graph Therapeutics brings total funding to over $10M to advance precision immunology — AI-powered analysis of live patient cells, multi-omics data, and functional biology decodes immune dysfunction for precision immunology.
• DeepRoot: A KG-Coordinated Multi-Agent System for Therapeutic Reasoning over Historical Medical Texts — DeepRoot's KG-coordinated multi-agent LLM system extracts drug leads from historical texts, separating grounding and reasoning for accuracy.
• GraphDiffMed: Knowledge-Constrained Differential Attention with Pharmacological Graph Priors for Medication Recommendation — Combining dual-scale differential attention with pharmacological graph priors significantly enhances medication recommendation safety and quality.
• TrialMatchAI: an end-to-end AI-powered clinical trial recommendation system to streamline patient-to-trial matching — TrialMatchAI automates clinical trial matching using LLMs and RAG for explainable, accurate patient eligibility.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Ai2.