About this project
Description from the program archive
This project will develop an intelligent chat interface for CellMinerCDB, a cancer genomics database, using GraphRAG (Graph Retrieval-Augmented Generation) to enhance accessibility for researchers. The system will convert natural language queries into executable R code and leverage knowledge graphs to provide context-aware responses about complex relationships between drugs, genes, and cancer cell lines.
Key deliverables include: (1) A data bridge connecting R's CellMinerCDB with Python; (2) A GraphRAG component that builds and queries knowledge graphs from cancer genomics data; (3) NDEx integration for network visualization and sharing; and (4) A natural language interface powered by PandasAI. This project extends previous work by adding knowledge graph capabilities, improving entity recognition for cancer genomics terminology, and enabling standardized network sharing through NDEx.
The completed system will democratize access to complex cancer genomics data, promote reproducibility in research, and establish a foundation for natural language interfaces to biomedical databases.