AI-enhanced adaptive virtual screening of large libraries for ligand discovery.

AdaptiveFlow enables ultralarge virtual screenings of 69 billion molecules, identifying nanomolar inhibitors for disease-relevant targets at unprecedented scale.

• Why it matters: Efficiently screening massive compound libraries is crucial for accelerating drug discovery but is limited by high costs and computational challenges, especially with growing chemical space.
• What they did: The platform integrates AI, ML, and over 1,500 docking protocols, utilizing active learning and cloud computing to reduce costs and scale up to millions of CPUs, while providing access to the largest drug-like molecule library.
• The result: AdaptiveFlow successfully discovered potent inhibitors of FSP1 and PARP1, offering mechanistic insights and enabling faster, more scalable development of AI-driven drug discovery methods.