Title: Machine learning-guided computational screening of new bio-orthogonal click reactions Show Chinese title

Title: 基于机器学习的新型生物正交点击反应计算筛选

Abstract: Bio-orthogonal click chemistry has become an indispensable part of the biochemist's toolbox. Despite the wide variety of applications that have been developed in recent years, only a limited number of bio-orthogonal click reactions have been discovered so far, most of them based on (substituted) azides. In this work, we present a computational workflow to discover new candidate bio-orthogonal click reactions. Sampling only around 0.05\% of an overall search space of over 10,000,000 dipolar cycloadditions, we develop a machine learning model able to predict DFT-computed activation and reaction energies within ~2-3 kcal/mol across the entire space. Applying this model to screen the full search space through iterative rounds of learning, we identify a broad pool of candidate reactions with rich structural diversity, which can be used as a starting point or source of inspiration for future experimental development of both azide-based and non-azide-based bio-orthogonal click reactions. Show Chinese abstract

Abstract: 生物正交点击化学已成为生物化学家工具箱中不可或缺的一部分。 尽管近年来已经开发了多种应用，但到目前为止发现的生物正偶性点击反应数量有限，其中大多数基于（取代的）叠氮化物。 在本工作中，我们提出了一种计算工作流程来发现新的候选生物正交点击反应。 仅对超过10,000,000种极性环加成的整体搜索空间进行约0.05%的采样，我们开发了一个机器学习模型，能够在整个空间内预测DFT计算的活化能和反应能，误差约为2-3 kcal/mol。 将该模型应用于整个搜索空间进行迭代学习，我们确定了一个具有丰富结构多样性的候选反应池，这可以作为未来基于叠氮化物和非叠氮化物的生物正交点击反应实验开发的起点或灵感来源。