St. Jude Children’s Research Hospital, University of Pavia, Dana-Farber Cancer Institute and Harvard Medical School unveiled AdaptiveFlow, an AI-informed platform claimed to virtually screen billions of drug-like molecules with around a 1,000-fold reduction in computational costs versus existing methods. The system was validated and published in Nature Biotechnology. The approach is positioned as an answer to the economics of ultra-large virtual screening—turning what is typically prohibitive compute into a workflow that teams can run routinely. If the claimed cost compression holds up across targets, it could accelerate hit discovery cycles and reduce time-to-signal in early-stage programs. For biotech development teams, the central question will be reproducibility across different protein targets and whether downstream hit triage improves clinical translation versus existing screening pipelines.
Get the Daily Brief