A conversation with Dr Rufus Daw about CellTypeAI. As single-cell datasets grow from thousands to hundreds of thousands of cells, accurate cell annotation is becoming a major bottleneck in biomedical research. In this episode, bioinformatician Rufus Daw discusses CellTypeAI, a new locally deployable AI framework that uses large language models to automate cell-type identification from single-cell RNA sequencing data. We explore how privacy-preserving AI tools, retrieval-augmented generation, and ensemble prompting can match or exceed conventional annotation approaches while keeping sensitive research data entirely in-house, and what this means for the future of AI-assisted immunology and systems biology.

 Further resources:


Rufus H. Daw,  Harry R. Deijnen, Magnus Rattray,  John R. Grainger, ‘CellTypeAI: Automated cell identification for scRNA-seq using local generative-AI’, preprinted posted on bioRxiv March 05, 2026.  https://doi.org/10.64898/2026.03.03.709253 [doi.org]
Rufus Daw talk: LLMs in the Laboratory: practical tools for AI-assisted research.

https://www.youtube.com/watch?v=I7AlwgYnu9c [youtube.com]