Automating Candidate Gene Prioritization with Large Language Models: From Naive Scoring to Literature-Grounded Validation Using Advanced NLP Technique

This dataset provides a persistent ChromaDB vector index built from a curated corpus of documents, designed for use in Retrieval-Augmented Generation (RAG) systems powered by large language models (LLMs). The index enables fast semantic retrieval and integrates seamlessly with the LlamaIndex framework. The index was generated using [SPECTRE2 , sentence-transformers/all-MiniLM-L6-v2] and is optimized for low-latency querying with applications in question answering, literature synthesis, and knowledge retrieval tasks. The full ChromaDB persistence directory is included, consisting of: chroma.sqlite — metadata and collection tracking collections/ — collection definitions index/ — serialized vector data This resource supports developers and researchers building domain-specific RAG pipelines, and can be used as a drop-in vector store within the LlamaIndex VectorStoreIndex loading pipeline. How to load: from llama_index.vector_stores.chroma import ChromaVectorStorefrom llama_index import StorageContext, load_index_from_storage persist_dir = "/path/to/unzipped_chromadb_index" vector_store = ChromaVectorStore(persist_dir=persist_dir)storage_context = StorageContext.from_defaults(persist_dir=persist_dir, vector_store=vector_store)index = load_index_from_storage(storage_context) 📌 Requires: llama-index, chromadb, sentence-transformers or custom SPECTER2 encoder if querying new docs. 🔬 Applications Scientific question answering Literature triage and summarization RAG-based expert assistant tools LLM-powered knowledge mining 🏷️ Keywords ChromaDB, LlamaIndex, SPECTER2, Scientific Retrieval, RAG, Vector Index, Semantic Search, LLM, Biomedical NLP, AI for Science