Discover Disease-Gene Associations, with Knowledge Graphs and AI

KG2ML-RAG is an integrated platform for exploring and prioritizing biomedical associations between diseases and entities such as genes and compounds. It uses Positive-Unlabeled (PU) learning to identify candidate associations that are not explicitly recorded in the Data Distillery Knowledge Graph (DDKG). Explore these associations through an interactive knowledge subgraph and use retrieval-augmented generation (RAG) to discover relevant scientific literature and generate evidence-based explanations, with citations to the underlying publications.