Microsoft Research and Partners Launch Project Quine, an AI System for Biological Research

On September 29, Microsoft Research, in collaboration with Harvard University and the Broad Institute of MIT and Harvard, unveiled Project Quine, an experimental multimodal AI research system. Designed as a "world model" for biological research, the system aims to bridge computational biology modeling with hands-on wet-lab experimentation.

Project Quine covers joint representations across genomics, proteins, chemistry, cellular states, and biological imaging, combined with an interactive reasoning framework. It can process information from multiple domains simultaneously for integrated cross-model analysis. Microsoft has opened applications for the inaugural Quine Fellows program and plans to gradually expand commercial access through products such as Microsoft Discovery.

Microsoft stressed that Project Quine is currently limited to research use and is not intended for clinical applications. Any results obtained with the system must undergo rigorous human evaluation. In a recent test on pancreatic ductal adenocarcinoma cell lines, the system prioritized compounds expected to drive cell state transition from classical to basal-like phenotype over a single weekend, and also surfaced unexpected phenotypic responses.

The system has the potential to shorten drug discovery timelines, saving years of work and millions of dollars, and could help researchers uncover previously unanticipated therapeutic pathways.