Researchers at the University of Cambridge have developed a new AI model capable of predicting the three-dimensional structures of proteins directly from their amino acid sequences with significantly improved speed and accuracy. The model, named AlphaFold 3, builds upon its predecessor by incorporating a diffusion-based architecture similar to those used in image generation. This allows it to generate more precise atomic-level structures, including how proteins interact with other molecules like DNA and drugs. The team reports that the model outperforms existing tools in the Critical Assessment of Protein Structure Prediction (CASP) competition, potentially accelerating drug discovery and fundamental biological research. The findings are detailed in the latest issue of the journal Nature. Read the full article at https://sciencedaily.com/releases/2024/05/240508113104.htm.