Enzyme activity from machine learning

Abstract

Enzyme Engineering Enzymes are very efficient catalysts for biochemical reactions, which are increasingly important for industrial applications. However, incomplete knowledge of the key factors that induce their catalytic properties limits our ability to engineer new enzymes with new properties. Bonk et al. propose a methodology for identifying those factors on the basis of a combination of machine-learning techniques and statistical mechanics methods. Their study of the rate-limiting step for an industrially important enzyme shows that conformational descriptors alone can be enough to predict reactivity and suggests that re-engineering enzymes to populate reactivity-promoting regions within the conformational space may lead to a significant improvement in catalytic performance. J. Am. Chem. Soc. 141 , 4108 (2019).