A multi-agent system for automating scientific discovery | Nature

Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain<br>the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in<br>Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles<br>and JavaScript.

Advertisement

A multi-agent system for automating scientific discovery

Download PDF

Subjects

Computational platforms and environments<br>Target identification

Abstract<br>Scientific discovery is driven by the iterative process of observation, hypothesis generation, experimentation, and data analysis. Despite recent advancements in applying artificial intelligence to biology, no system has yet automated all these stages [1, 2, 3]. Here, we introduce Robin, the first multi-agent system capable of fully automating both hypothesis generation and data analysis for experimental biology. By integrating literature search agents with data analysis agents, Robin can generate hypotheses, propose experiments, interpret experimental results, and generate updated hypotheses, achieving a semi-autonomous approach to scientific discovery. By applying this system, we were able to identify promising therapeutic candidates for dry age-related macular degeneration (dAMD), the major cause of blindness in the developed world [4, 5]. Robin proposed enhancing retinal pigment epithelium phagocytosis as a therapeutic strategy, and identified and confirmed in vitro efficacy for ripasudil and KL001. Ripasudil is a clinically-used Rho kinase (ROCK) inhibitor that has never previously been proposed for treating dAMD. To elucidate the mechanism of ripasudil-induced upregulation of phagocytosis, Robin then proposed and analyzed a follow-up RNA-seq experiment, which revealed upregulation of ABCA1, a lipid efflux pump and possible novel target. All hypotheses, experimental directions, data analyses, and data figures in the main text of this report were produced by Robin. As the first AI system to autonomously discover and validate novel therapeutic candidates within an iterative lab-in-the-loop framework, Robin establishes a new paradigm for AI-driven scientific discovery.

You have full access to this article via your institution.

Download PDF

Similar content being viewed by others

Discovery of a selective alpha-kinase 1 inhibitor for the rare genetic disease ROSAH syndrome

Article<br>Open access<br>09 September 2025

Integrating machine learning and structure-based approaches for repurposing potent tyrosine protein kinase Src inhibitors to treat inflammatory disorders

Article<br>Open access<br>13 January 2025

Effect of a fixed combination of ripasudil and brimonidine on aqueous humor dynamics in mice

Article<br>Open access<br>03 April 2024

Author information<br>Author notesThese authors contributed equally: Ali Essam Ghareeb, Benjamin Chang

These authors jointly supervised this work: Andrew D. White, Michaela M. Hinks, Samuel G. Rodriques

Authors and Affiliations<br>FutureHouse, San Francisco, USA<br>Ali Essam Ghareeb, Benjamin Chang, Ludovico Mitchener, Angela Yiu, Caralyn J. Szostkiewicz, Jon M. Laurent, Samantha M. Wright, Muhammed T. Razzak, Andrew D. White, Silvia C. Finnemann, Michaela M. Hinks & Samuel G. Rodriques

University of Oxford, Oxford, UK<br>Benjamin Chang

Fordham University, New York, USA<br>Dmytro Shved & Gavin J. Gyimesi

AuthorsAli Essam GhareebView author publications<br>Search author on:PubMed Google Scholar

Benjamin ChangView author publications<br>Search author on:PubMed Google Scholar

Ludovico MitchenerView author publications<br>Search author on:PubMed Google Scholar

Angela YiuView author publications<br>Search author on:PubMed Google Scholar

Caralyn J. SzostkiewiczView author publications<br>Search author on:PubMed Google Scholar

Dmytro ShvedView author publications<br>Search author on:PubMed Google Scholar

Gavin J. GyimesiView author publications<br>Search author on:PubMed Google Scholar

Jon M. LaurentView author publications<br>Search author on:PubMed Google Scholar

Samantha M. WrightView author publications<br>Search author on:PubMed Google Scholar

Muhammed T. RazzakView author publications<br>Search author on:PubMed Google Scholar

Andrew D. WhiteView author publications<br>Search author on:PubMed Google Scholar

Silvia C. FinnemannView author publications<br>Search author on:PubMed Google Scholar

Michaela M. HinksView author publications<br>Search author on:PubMed Google Scholar

Samuel G. RodriquesView author publications<br>Search author on:PubMed Google Scholar

Corresponding authors<br>Correspondence to<br>Andrew D. White, Michaela M. Hinks or Samuel G. Rodriques.

Supplementary information

Supplementary Information (download PDF )<br>Combined PDF containing Supplementary Tables 1-2, Supplementary Figures 1-20, and Supplementary References. The tables list drugs with working concentrations and antibodies used in this paper. The figures detail...