Bing:
- Researchers at MIT have developed a generative AI system that can design new protein structures from scratch.
- The system, called ProteinGAN, uses a deep neural network to learn from existing protein data and generate novel 3D shapes that could have biological functions.
- ProteinGAN could help scientists discover new drugs, enzymes, and biomaterials by expanding the diversity and complexity of protein designs.
- The system is also able to generate protein structures that are similar to those of a given target, such as a virus or a receptor, which could aid in designing inhibitors or agonists.
- The researchers tested ProteinGAN on several benchmarks and found that it outperformed existing methods in terms of novelty, diversity, and quality of the generated proteins.