(27ab) Enzyme Function Prediction Using Contrastive Learning

Enzyme function annotation is a fundamental challenge and numerous computational tools have been developed. However, most of these tools cannot accurately predict functional annotations such as enzyme commission (EC) number for less studied or proteins with novel or multiple functions. Herein, we present a machine learning (ML) algorithm named CLEAN (contrastive learning enabled enzyme annotation) to assign EC numbers to enzymes with better accuracy, reliability, and sensitivity than the state-of-the-art tool BLASTp. The contrastive learning framework empowers CLEAN to confidently: i) annotate the understudied enzymes, ii) correct the mislabeled enzymes, and iii) identify the promiscuous enzymes with two or more EC numbers, which are demonstrated by systematic in silico and in vitro experiments. We expect this tool to greatly facilitate enzymology and synthetic biology studies. It is worth mentioning that this work was recently published in Science (Yu et al. Science 379, 1358–1363 (2023)) and received much attention from the broad research community.