Chiral alcohols, as essential building blocks for active pharmaceutical ingredients and fine chemicals, have found broad applications in pharmaceuticals, agrochemicals, and chemical industries. However, conventional synthetic strategies often suffer from low efficiency and insufficient selectivity, highlighting the urgent need for innovative catalytic systems to achieve efficient asymmetric synthesis. Alcohol dehydrogenases (ADHs) and carbonyl reductases (CRs) are key biocatalysts enabling the green synthesis of chiral alcohols, and their molecular engineering and industrial deployment have emerged as research frontiers. Despite their intrinsic advantages of environmental compatibility and catalytic specificity, the practical application of ADHs/CRs remains hindered by their limited substrate promiscuity and poor substrate adaptability, which severely restrict large-scale implementation in the precise synthesis of chiral alcohols. This review provides a systematic overview of recent advances in quantitative multisubstrate fitness studies of ADHs/CRs, with a particular emphasis on methodologies for elucidating the quantitative structure-activity relationships (QSARs) across ADH libraries and diverse substrate panels. The perspectives presented aim to advance fundamental understanding of multi-enzyme-multi-substrate QSARs, enabling the development of intelligent design platform trained on tens of millions of QSAR data points. Such innovations promise to revolutionize alcohol dehydrogenase engineering paradigms, shifting the paradigm for optimal enzyme discovery in chiral alcohol synthesis from “needle-in-a-haystack” screening to “tailor-made” precision design.

Key words: alcohol dehydrogenase/carbonyl reductase, substrate library construction, enzyme library construction, quantitative structure-activity relationship, biocatalysis database, deep learning