The Pulse
Microsoft Says Quine Narrowed a Cancer-Drug Search in One Weekend
Microsoft Research introduced Quine, an AI system that connects biological data, research tools and lab experiments. In a pancreatic cancer study with the Broad Institute, Microsoft says Quine ranked thousands of compounds and helped priori
AI.info Team · Written with AI ·
One weekend was enough for Microsoft Research’s Quine system to narrow a search across thousands of compounds and prioritize candidates for lab testing, the company said on September 29. In work with researchers at the Broad Institute of MIT and Harvard, Microsoft reports that the highest-ranked compounds produced the largest intended shifts in pancreatic cancer cells across experimental assays. The result is an early demonstration of Quine’s central idea: use a computational model to choose experiments, then feed the measurements back into the next round of research.
Quine links biological data to lab tests
Microsoft describes Quine as a research system combining a “world model” of biology with an interactive harness that connects models to scientific tools, literature, researchers and wet-lab work. Its model learns across biological data types and scales, including genomic sequences, proteins, chemical compounds, cellular state and imaging. The aim is to let evidence from one type of data inform predictions about another, rather than relying only on separate specialist models for each task.
The system is designed to represent biological states, predict how interventions could change them and help researchers compare possible experiments before spending time and resources in the lab. Microsoft says Quine does not need to model biology perfectly to be useful; it needs to help scientists design better experiments. The company frames the work as an iterative loop: researchers ask questions, Quine proposes and ranks possibilities, experiments test them, and results inform what comes next.
Thousands of compounds, one reported weekend
The cancer example focuses on pancreatic ductal adenocarcinoma, or PDAC, and whether tumor cells’ transcriptional states could be actionable targets. The team examined a transition from a “classical” to a “basal” cell state, using Quine to predict and prioritize thousands of compounds according to their potential to shift the cells. Microsoft says the highest-ranked candidates produced the strongest intended shifts in wet-lab assays.
“The entire process—from rapidly narrowing the compound search space to prioritizing a handful of promising candidates to be validated in the lab—took just one weekend, potentially saving months of experimental work and significant research costs.”
Nicolo Fusi, VP and Distinguished Scientist, and Jonathan M. Carlson, Vice President, Microsoft Research
The reverse transition, from basal back to classical, proved harder, and Quine predicted a weaker effect from available compounds. Experiments also supported the model’s prediction that several compounds would move cells toward a third phenotype, rather than along a simple two-state axis. Microsoft says that finding points to more complexity in the cell behavior under study and plans to incorporate additional RNA data to improve its predictions.
Promising results, with limits
The announcement offers a concrete example, but it does not name the compounds or provide enough detail in the post to assess their individual effects, reproduce the analysis or compare Quine’s performance with other methods. The reported cell-state shifts are experimental findings, not evidence that the compounds treat cancer in patients. Microsoft says some of the stronger effects came from compounds with unexpected mechanisms, presenting them as possible leads for drug repurposing and discovery—not established therapies.
Microsoft also stresses that Quine is experimental research technology, not for clinical or medical use. Its outputs may be incomplete or inaccurate and require review by qualified researchers and scientific validation. Access is initially limited to the Quine Fellows program and selected research collaborations; applications for the first cohort are open, according to the announcement.
Scientists get the first chance to test it
Microsoft says Quine grew out of its own research programs in cancer biology, protein engineering, genomics and bioimaging, and that it wants outside scientists to help shape the system. The company expects access to expand through products such as Microsoft Discovery as the technology matures, but gives no timetable. The next test is whether independent researchers can reproduce useful results across other biological questions—and whether Quine’s ranked candidates continue to perform when put through further experiments.