AUTOMATED SINGLE-CELL RNA SEQUENCING ANALYSIS SUPPORTED BY LARGE LANGUAGE MODELS

Data-driven methods, including machine learning, have advanced single-cell RNA sequencing (scRNA-seq) analysis, but they often fail to integrate research context, limiting their ability to provide in-depth, contextual insights to researchers. We introduce scChat, an AI-powered assistant designed for contextualized scRNA-seq analysis. Unlike existing tools, scChat enables research context-driven interpretation, hypothesis validation, and experimental planning. By integrating large language models (LLMs), function calls for quantitative analysis, retrieval-augmented generation to reduce hallucinations, and an LLM-powered search engine for literature integration, scChat enhances the accuracy and depth of scRNA-seq insights. Through case studies on glioblastoma datasets, scChat effectively identifies cell populations, explains immune dynamics, and suggests mechanistic insights, demonstrating its potential as an interactive, AI-driven co-pilot for single-cell transcriptomics research.