Google DeepMind has announced a major advance in computational biology with a new AI system capable of predicting protein-protein interactions with unprecedented accuracy. The system, an evolution of the AlphaFold architecture, can model how proteins bind to each other and to small molecules, opening new possibilities for drug discovery and disease understanding.
In benchmark tests, the system correctly predicted interaction patterns for 94 percent of known protein complexes, compared to 67 percent for the previous best computational method. Researchers say the tool could dramatically accelerate the identification of drug targets and the design of therapeutic molecules.
The technology is being made available to academic researchers through a partnership with the European Molecular Biology Laboratory. Pharmaceutical companies, including Roche and Novartis, have already signed licensing agreements to integrate the system into their drug discovery pipelines. DeepMind says the advance represents a step toward a comprehensive computational model of cellular biology.
