CellVoyager is an innovative AI agent designed to automate the analysis of complex single-cell RNA sequencing (scRNA-seq) data. By utilizing large language models like o3-mini, the system autonomously generates hypotheses, writes executable Python code, and interprets results within Jupyter notebooks. It specifically aims to complement human research by identifying novel biological insights that may have been overlooked in original studies. Evaluated against a new benchmark called CellBench, the agent outperformed standard models in predicting relevant analytical workflows. Practical case studies in COVID-19, aging, and cell communication demonstrated its ability to uncover scientifically sound findings that experts rated as highly creative. This framework represents a significant step toward autonomous computational biology, allowing for the rapid reanalysis of vast public datasets.

References:

• Alber S, Chen B, Sun E, et al. Cellvoyager: Ai compbio agent generates new insights by autonomously analyzing biological data[J]. Nature Methods, 2026: 1-11.