pipette
ESEspañol

An open agentic AI co-scientist for clinical and biomedical discovery

Kaleigh F. Roberts, Zachary B. Abrams, Luca Cappelletti, Mahdi Moqri, Nicholas Heugel, John Harry Caufield, Mathieu Bourdenx, Yan Li, Jineta Banerjee, Luca Foschini, Diego Galeano, Nomi L. Harris, Melody Li, Kejun Ying, Justin Melendez, Nicolas R. Barthélemy, James G. Bollinger, Yingxin He, Vitaliy Ovod, Tammie L.S. Benzinger, Shaney E. Flores, Brian Andrew Gordon, Taylor Pedersen, Jinsheng Yu, Brendan P. Lucey, Erik S. Musiek, Sarah A. Naylor, Adegoke A. Ojewole, Mukta Phatak, Donald Leo Elbert, Sarah Biber, Eric C. Landsness, Chris Mungall, Randall John Bateman, Justin Reese

Peer-reviewed journalBold claims, read criticallyReal-world use

In the authors' words

Abstract Biomedical research is constrained by the time and domain expertise required to analyze increasingly large and complex datasets. Here we present OpenScientist, an open-source agentic AI co-scientist designed to address this gap by investigating scientist-defined queries and generating verifiable, data-grounded analyses. OpenScientist was evaluated by domain experts across five biomedical case studies, completing analyses in minutes that otherwise would take months of human time and expertise. It accurately quantified Alzheimer’s blood biomarker performance, generated a plasma proteomic survival model that predicted survival in a hidden-evaluation cohort, proposed a reproducible mechanism linking tau pathology to altered lysosomal acidification, generated multiple myeloma hypotheses validated in an external cohort with attenuated signal in randomized controls, and demonstrated concordance with an independent human analysis in a blinded, prespecified proteomic comparison. These results support the use of OpenScientist to assist biomedical data analysis in concert with human experts. To support research involving sensitive clinical datasets, OpenScientist can operate in a fully air-gapped environment using open-source models for inference. OpenScientist demonstrates that open, auditable, agentic AI can support real-world clinical research by researching the literature, generating hypotheses, executing analyses, and generating insights from complex datasets.

Main resultThe abstract does not state a limitation.

Appeared: Wednesday, November 18. Nature Medicine. Peer-reviewed journal.

DOI: 10.1038/s41591-026-04808-6