High-throughput Prediction of Enzyme Promiscuity Based on Substrate–Product Pairs: An enzyme screening tool

Enzyme screening is an essential preliminary activity in metabolic engineering. Nonetheless, the existing tools largely rely on prior knowledge, and cannot utilize custom candidate enzyme libraries. To address this, we introduced the Substrate–product Pair-based Enzyme Promiscuity Prediction (SPEPP) model, which leverages transfer learning and Transformer architecture to illuminate the intricate interplay between enzymes and substrate–product pairs. SPEPP exhibited good predictive ability, eliminating the need for prior knowledge of reactions and allowing users to define their candidate enzyme libraries. Owing to its adaptability, SPEPP can be seamlessly integrated into various metabolic engineering applications including, but not limited to, substrate/product screening, de novo pathway design, and hazardous material degradation. To better assist metabolic engineers in designing and refining biochemical pathways, particularly those without programming skills, we designed EnzyPick, an easy-to-use web server for enzyme screening, based on SPEPP.