目的 针对中药（traditional Chinese medicine，TCM）分子结构复杂、环系结构多样等特点，优化图同构神经网络（graph isomorphism network，GIN），提升其对中药分子关键性质的预测能力。方法 基于GIN模型，在分子图中引入表示键长张力和角度张力的边描述符，并结合多个优势构象构建多图数据，从而构建了一种柔性-环增强图神经网络（flexibility-ring enhanced graph neural network，FRGNN）。在包含37 822个分子的2个中药数据库中，评估了该模型在7项关键分子属性上的预测性能，并与3种当前最优（state-of-the-art，SOTA）图神经网络模型及2种基础图神经网络模型进行了对比。结果 与表现次优的模型相比，FRGNN在7项性质预测上的均方根误差平均降低了8.63%；对于含多环及大环结构的分子，其预测均方根误差进一步降低10.04%。结论 FRGNN模型在中药分子关键性质预测任务中表现出优于现有小分子预测SOTA模型的性能，为复杂天然化合物的性质预测提供了新的有效方法。

Objective To address the complex molecular structures and diverse ring systems characteristic of traditional Chinese medicine (TCM) compounds, this study aims to optimize the graph isomorphism network (GIN) to enhance its predictive capability for key molecular properties of TCM molecules. Methods Based on the GIN architecture, edge descriptors representing bond length strain and angle strain were introduced into the molecular graph, and multi-graph data were constructed by incorporating multiple favorable conformations, thereby developing a flexibility-ring enhanced graph neural network (FRGNN). The predictive performance of the proposed model on seven key molecular properties was evaluated using two TCM databases containing 37 822 molecules, and compared against three state-of-the-art (SOTA) graph neural network models and two basic graph neural network models. Results Compared with the second-best performing model, FRGNN achieved an average reduction of 8.63% in root mean square error across the seven property prediction tasks, for molecules containing polycyclic and macrocyclic structures, the predictive root mean square error was further reduced by 10.04%. Conclusion The proposed FRGNN model demonstrates superior performance over existing state-of-the-art models for small-molecule property prediction in the task of predicting key properties of TCM compounds, providing a novel and effective approach for the property prediction of complex natural products.

R284.1

国家自然科学基金面上项目（22578039）；重庆市璧山区中医药领航联合专项重大项目（BSZYYLH001）；重庆市自然科学基金创新发展联合基金（CSTB2023NSCQ-LZX0060）