URS Protein Mutation Prediction — ClinVar and UniProt RAG Vector Stores

Pre-built ChromaDB vector databases for the URS Protein Mutation Prediction pipeline. Two stores are included: chroma_clinvar/ — ClinVar variant embeddings for semantic and keyword retrieval of clinical pathogenicity classifications (Pathogenic, Benign, VUS, etc.) chroma_uniprot/ — Curated Swiss-Prot/UniProt embeddings covering protein identity, function, domain structure, GO terms, and mutagenesis evidence These stores enable retrieval-augmented generation (RAG) without rebuilding the databases from scratch. Download both folders and place them as described in the repository README. Requires chromadb (installed automatically via pip install -r requirements.txt).