Generative artificial intelligence for enzyme design: Recent advances in models and applications
Shuixiu Wen, Wen Zheng, Uwe T. Bornscheuer, Shuke Wu
Abstract
Enzyme catalysis is a key enabling technology for green and sustainable production of chemicals. Developing suitable enzymes is at the heart of this technology, which is currently changing by Artificial Intelligence (AI) such as machine learning. AI-based methods were used for enzyme discovery and design. We review the recent advances in generative AI models for enzyme design, with a particular focus on those that have been validated by experiments. Furthermore, we discuss the applications of the enzymes designed by generative AI, including artificial luciferases , non-heme iron (II)-dependent oxygenases , and P450 enzymes. We provide our opinions on several current issues encountered in computational enzyme design. With the fast development of new generative models in enzymes and the implementation of these models by the research community, we believe that the precise design of efficient enzymes with new catalytic functions and/or potential industrial applications will be a mature method in the near future.