AI structure prediction speeds discovery of 'molecular glues' to treat disease
MedicineChemistryBiology
THE AI ANGLE
Predicting three-part protein complexes without prior experimental structuresResearchers combined high-throughput proteomics with computational modeling to discover molecular glues that direct the VAV1 protein to cellular disposal machinery. The platform predicted how the molecules bind target proteins without prior experimental structures, which helped chemists design more potent variants earlier in the process.
THE TEACHING ANGLE
Drug developers must balance the speed of predictive modeling with the risk of off-target effects, such as the unintended degradation of LIMD1.Read the original at phys.org Generate teaching or study materials
More in Medicine
- Drones, satellites and AI help scientists watch over Antarctica's smallest lifeformsPhys.org — Biology · September 30, 2026
- Unlocking AI as a research tool for plant microscopyPhys.org — Biology · September 30, 2026
- Physics-grounded AI framework aims to make predictions about new materials more testablePhys.org — Chemistry · September 30, 2026
- Wearable AI forecasts prolonged sitting in women with chronic pelvic painMedical Xpress · September 30, 2026
- Machine learning toolkit can improve odds of drug development successMedical Xpress · September 30, 2026