Megalibrary platform rapidly discovers promising new materials. Platform marks a shift from trial-and-error materials discovery toward intentional design. Team used the system to identify a new piezoelectric material and then deliberately engineered it to operate at a target temperature. Megalibraries outperform emerging ‘self-driving labs’ in terms of speed and throughput to create and screen vast numbers of materials samples simultaneously.

Researchers publishing in Computational Biomedicine have developed a novel artificial intelligence framework, MIGDTA, that integrates graph neural networks and multi-source biological information to enhance drug–target affinity prediction. The study highlights how combining molecular structure, biological descriptors, and interaction networks could accelerate drug discovery and repositioning.