Proto-OKN Masterclass: Building and Querying the SPOKE-GeneLab Knowledge Graph

Описание: Welcome to the Proto-OKN Masterclass Series!

In this episode, Peter Rose presents a technical overview of building and querying the NASA SPOKE–GeneLab Knowledge Graph, which integrates space biology data with comprehensive biomedical knowledge.

The talk introduces a Model Context Protocol (MCP) server (mcp-genelab) that translates natural language questions into Cypher queries executed against Neo4j endpoints. Custom tools enable AI-assisted access to NASA GeneLab data, supporting analysis of differential gene expression and DNA methylation from spaceflight experiments via conversational AI interfaces.

By integrating data from NASA’s GeneLab Data Repository, part of the NASA Open Science Data Repository (OSDR), with the Scalable Precision Medicine Open Knowledge Engine (SPOKE) knowledge graph, the system connects spaceflight results to rich biological context, including genes, anatomy, and disease.

The session includes schema design, query examples, and live demonstrations.

Presentation: https://nebigdatahub.org/wp-content/u...
Installation, code, and examples: https://github.com/sbl-sdsc/mcp-genelab
NASA SPOKE-Genelab Knowledge Graph: https://github.com/BaranziniLab/spoke...
Questions & issues: https://github.com/sbl-sdsc/mcp-genel...

Follow along with the Proto-OKN Masterclass Series to learn more about the collaborative projects that make up the Proto-OKN.

_________________________________________________________________________

Learn more about the Proto-OKN:
In collaboration with five other U.S. government agencies, the U.S. National Science Foundation (NSF) has invested $26.7 million in 18 projects through its Building the Prototype Open Knowledge Network (Proto-OKN) program. An open knowledge network is a publicly accessible, interconnected set of data repositories and associated knowledge graphs that will enable data-driven, artificial intelligence-based solutions for a broad set of societal and economic challenges.