Project
PaA-10
AI-to-Action Drug Design: Generative Selection with Industry Validation
Project Heads
Konstantin Fackeldey, Christof Schütte
Project Members
Christopher Secker
Project Duration
01.09.2026 − 31.08.2028
Located at
ZIB
Description
Generative Selection in Drug Design turns discovery into a closed loop: start with a diverse pool, let generative AI propose smarter molecules, validate them with high-throughput screening plus additional metrics for pharmacokinetic and safety, then keep only the elites and learn from them to propose even better candidates next round. Built on high-throughput pipelines like VirtualFlow, the project adds new mathematics to guarantee stability and convergence of the loop. Proven on pH-selective pain-relief leads and viral protease inhibitors, and next applied to drug development in collaboration with an industry partner, the goal is to scale with pharma and biotech partners towards a spin-off company.
Project Webpages
Related Publications
Related Picture