A specialized assistant in Versa that enhances Large Language Models with SPOKE biomedical knowledge graph
Quick Start β’ KG-RAG β’ Examples β’ API Documentation
- Quick Start
- What is SPOKE?
- What is KG-RAG?
- Using KG-RAG in UCSF Versa
- KG-RAG API Access
- Additional Resources
- Access Versa (requires UCSF authentication)
- Select "SPOKE - Knowledge Graph" from Assistants dropdown
- Choose your preferred language model (e.g., GPT-4o)
- Ask your disease-related biomedical question
- Review the evidence-based response with sources
SPOKE (Scalable Precision Medicine Open Knowledge Engine) is UCSF's comprehensive biomedical knowledge graph that integrates and connects information from over 40 specialized databases. It serves as a unified platform for biomedical knowledge, making complex relationships between different biological entities discoverable and accessible.
- Nodes: 27+ million nodes of 21 different types
- Edges: 53+ million edges of 55 different types
- Sources: Integrates 41+ specialized biomedical databases
- Updates: Refreshed weekly to ensure current information
- Prioritizes experimentally validated information
- Maintains clear provenance for all relationships
- Provides statistical evidence when available (p-values, z-score, confidence scores)
- Focuses on curated databases rather than text mining
- Disease-Disease Ontology associations
- Disease-Gene associations
- Disease-Symptom relationships
- Disease-Compound associations
- Compound-Protein interactions
- Protein-Protein interactions
- Anatomical hierarchies and many more
π For more detailed information about SPOKE, visit the SPOKE Explorer or read the SPOKE Paper.
KG-RAG (Knowledge Graph Retrieval Augmented Generation) is a specialized framework that enhances Large Language Models (LLMs) with SPOKE's biomedical knowledge. By combining the reasoning capabilities of LLMs with verified biomedical information from SPOKE, KG-RAG provides reliable, evidence-based responses to biomedical questions.
-
Large Language Models (LLMs)
- Advanced AI models like GPT-4
- Natural language understanding and generation
- Reasoning capabilities
-
SPOKE Knowledge Graph
- Verified biomedical knowledge
- Structured relationships
- Statistical evidence and provenance
-
Sentence Transformers
- Creates embeddings of biomedical context from SPOKE
- Creates embeddings of user queries
- Enables context pruning through semantic similarity, thereby optimizes the knowledge retrieval
-
Disease Recognition
- Uses sentence transformers to embed user questions
- Matches with disease entities in SPOKE
- Ensures accurate disease identification
-
Context Retrieval & Pruning
- Extracts relevant context from SPOKE
- Uses sentence transformers to embed biomedical context
- Prunes context based on semantic similarity to query
-
Context Enhancement
- Combines pruned knowledge with LLM capabilities
- Pruning optimizes the token utilization for the LLM
- Preserves evidence and provenance information
-
Response Generation
- Generates comprehensive answers grounded on factual biomedical knowledge from SPOKE
- Includes evidence-based support
- Maintains scientific accuracy
- Provided provenance in the generated text
-
Verified Information
- Responses backed by SPOKE's curated biomedical knowledge
- Clear provenance for all information
-
Scientific Accuracy
- Statistical evidence when available (Note: Versa maynot support this, but KG-RAG API does)
- Multiple source validation
-
Intelligent Disease Recognition
- Robust entity recognition using embeddings
- Handles variations in disease names
- Maps to standardized disease concepts
-
Smart Context Retrieval
- Semantic matching for relevant information
- Efficient pruning of knowledge graph data
- Optimal context selection
-
Token Efficiency
- Optimized context selection
- Reduced token usage compared to traditional RAG
- Cost-effective implementation
-
Consistent Results
- Stable responses across LLM updates
- Evidence-based conclusions
Examples of KG-RAG's comprehensive responses with statistical evidence. Note that these queries were run in March 2024 using GPT-4. When running the same queries today, GPT-4-only responses (blue box) may differ from those shown in the figure due to the possible model updates from OpenAI.
Query: "Are there any latest drugs used for weight management in patients with Bardet-Biedl Syndrome?"
KG-RAG provides:
- Retrieves drug treatment information from SPOKE
- Shows clinical trial phase (Phase 3)
- Includes multiple source databases (ChEMBL, DrugCentral)
- Provides clear provenance for information
Query: "Is it PNPLA3 or HLA-B that has a significant association with the disease liver benign neoplasm?"
KG-RAG provides:
- Comparative statistical analysis
- Precise p-values (PNPLA3: 4e-14, HLA-B: 2e-08)
- Source attribution (GWAS Catalog)
- Evidence-based conclusion
π‘ Note: These examples demonstrate KG-RAG's full capabilities with statistical evidence. The current Versa implementation may have different features, which we'll discuss in the Using KG-RAG in Versa section.
π For more technical details about KG-RAG's architecture and performance, read our research paper published in Bioinformatics.
KG-RAG is available in UCSF Versa as a specialized assistant that enables biomedical question-answering using SPOKE knowledge. Here's how to effectively use KG-RAG in Versa:
-
Connect to UCSF VPN To access the Versa application, users must first connect to the UCSF VPN.
-
Select the Assistant
- Choose "SPOKE - Knowledge Graph" from the Assistants dropdown menu of Versa
- Select your preferred language model (e.g., GPT-4o)
- Frame Your Question
- Currently, Versa accepts only disease-related queries (i.e. queries that have disease names mentioned in it. e.g. what are the genes associated with multiple sclerosis?)
- Be specific and clear in your questions
π‘ Disease Coverage: SPOKE contains 11,697 disease concepts, providing comprehensive coverage across various medical domains. This means users can inquire about a wide spectrum of diseases, from common conditions to rare disorders, all backed by verified biomedical knowledge.
- For now, Versa's KG-RAG implementation focuses on disease-centric questions
- Responses will include information sourced from SPOKE
- While statistical evidence is available in SPOKE, the current Versa implementation doesn't use that (you can get that information using KG-RAG API)
KG-RAG in Versa provides structured responses in three sections:
-
SPOKE-Prioritized Response
- Presents findings directly from SPOKE knowledge base
- Lists entities with their relationships
- Includes provenance information (data sources)
- Complemented with relevant LLM knowledge
-
Analysis Without SPOKE
- Provides context from LLM's training
- Offers additional insights
- Helps validate and complement SPOKE information
-
Summary
- Combines insights from both sources
- Highlights key findings from SPOKE
- Includes additional context from LLM
- Provides comprehensive conclusions
π‘ Example Response Format:
SECTION 1 - SPOKE-PRIORITIZED RESPONSE:
Based on SPOKE knowledge base:
* Information directly from SPOKE with provenance
* Additional context from LLM
SECTION 2 - ANALYSIS WITHOUT SPOKE:
Based on trained biomedical knowledge:
* LLM's knowledge about the topic without using SPOKE
SECTION 3 - SUMMARY:
From Section 1 (with SPOKE):
* Key findings from SPOKE
From Section 2 (without SPOKE):
* Key findings from LLM
Final comprehensive conclusion combining insights from Section 1 and Section 2.
Based on SPOKE's knowledge graph structure, you can ask questions about:
π Gene-Disease Associations Example: "What genes are associated with Acute Monocytic Leukemia?"
π Disease-Disease Similarity Example: "Which diseases are similar to Parkinson's disease?"
π Disease-Disease Ontology Example: "What is the disease ontology of Alzheimer's disease?"
π Disease-Drug Treatments Example: "What drugs are used to treat multiple sclerosis?"
π Disease-Symptom Relationships Example: "What are the symptoms of Bardet-Biedl Syndrome?"
π Disease-Organism Associations Example: "Which organisms can cause pneumonia?"
π Disease-Anatomy Localization Example: "Which anatomical structures are affected by diabetes?"
KG-RAG also supports queries that combine multiple relationship types or explore intersections between two or more diseases. Here are some examples:
π Disease-Gene-Disease Connections Example: "What genes are common between Parkinson's disease and Alzheimer's disease?"
π Disease-Symptom-Disease Patterns Example: "What symptoms are shared between multiple sclerosis and lupus?"
π Disease-Drug-Disease Relationships Example: "What drugs are used to treat both rheumatoid arthritis and psoriatic arthritis?"
π Disease-Anatomy-Disease Associations Example: "Which anatomical structures are affected by both diabetes and hypertension?"
π‘ Tip: When forming intersection queries, clearly specify both diseases and the relationship type you're interested in exploring between them.
Current average response time:
- GPT-4o: 24.5 Β± 17.7 seconds
- GPT-4: 30.3 Β± 9.6 seconds
These latencies are due to multiple API calls in the backend pipeline:
- User query β GPT API (disease entity extraction)
- Azure API (semantic search)
- KG-RAG API (context extraction from SPOKE)
- GPT API (response generation and summarization)
- Currently limited to disease-centric questions
- Queries must explicitly mention disease names
- Other biomedical queries (e.g., drug-protein interactions without disease context) are not supported in the current Versa implementation
- For broader biomedical queries, consider using the KG-RAG API
- Versa's implementation uses single-hop graph search for optimal performance
- While deeper graph searches are possible through the KG-RAG API, they result in:
- Exponential increase in response time
- Larger context volume
- Higher API costs
For a comprehensive collection of example queries and their responses, visit our Examples Guide.
While KG-RAG is integrated into Versa for disease-centric queries, you can also access it directly through our REST APIs for broader biomedical questions. We provide two specialized endpoints:
-
Disease-Centric Endpoint (
v1/kg_rag_context
)- Optimized for disease-related queries
- Used by Versa integration
- Requires specific disease nodes
-
Extended Endpoint (
v1/kg_rag_context_extended
)- Supports broader biomedical queries
- No disease node requirement
- More flexible querying capabilities
- Access to complete SPOKE knowledge
- Statistical evidence inclusion option
- Configurable search depth
- Detailed provenance information
For detailed documentation, including:
- Complete API reference
- Code examples
- Response formats
- Implementation guidelines
π Please refer to our API Documentation
- KG-RAG Paper - Technical details about KG-RAG framework and its performance
- SPOKE Paper - Comprehensive overview of SPOKE knowledge graph
- KG-RAG GitHub Repository - Open-source code and setup instructions
- SPOKE API Documentation - Complete SPOKE API reference
- SPOKE Explorer - Interactive interface to explore SPOKE knowledge graph
π‘ Want to run KG-RAG on your machine? Follow the instructions in the KG-RAG GitHub repository to set up and run KG-RAG locally.