中草药

基于分子结构跨域泛化的中药成分相互作用预测研究

PDF下载

程宁, 曾湘祥, 宋学坤

中草药 | 数据挖掘与循证医学 2026,57(17): 6867-6877

收起

中草药 |数据挖掘与循证医学 2026 , 57 (17) : 6867 -6877

基于分子结构跨域泛化的中药成分相互作用预测研究

全屏

程宁, 曾湘祥, 宋学坤

作者信息

通讯作者:

宋学坤

作者简介:

程宁: 程宁，讲师，主要从事中医药人工智能研究。E-mail:chengning@hactcm.edu.cn

Prediction of traditional Chinese medicine component interactions via cross-domain generalization of molecular structure

CHENG Ning, ZENG Xiangxiang, SONG Xuekun

Affiliations

doi: 10.7501/j.issn.0253-2670.2026.17.018

文章导航

摘要

收起

目的针对中药复方成分复杂、相互作用关系不明导致配伍机制难以解析的问题，以成分间相互作用预测为切入点，构建基于深度学习的分子相互作用预测模型，为识别中药成分间潜在相互作用关系及后续方剂配伍机制研究提供方法学参考。方法提出基于分子3D结构特征与扩展连通性指纹的分子相互作用预测模型（3D structure and extended connectivity fingerprints molecule-molecule interaction，3DFPMMI）。模型利用SchNet网络提取分子3D图结构的深层特征，同步获取扩展连通性指纹，以对比学习策略实现多源特征的有效融合；模型在DrugBank药物相互作用数据集上完成训练，并通过跨域泛化实现中药成分间相互作用的预测。结果 3DFPMMI模型在DrugBank数据集上准确率达到0.921，精确率-召回率曲线下面积为0.972，均优于其他基准模型。消融实验证实，分子3D结构特征、扩展连通性指纹特征及对比学习特征融合策略均为模型性能提升的核心有效模块。将该模型应用于经典方剂补阳还五汤和半夏泻心汤的成分相互作用预测，其部分预测结果与已有的实验研究结论高度吻合。结论构建了一种中药成分相互作用预测的新方法，并初步验证了其应用潜力，为从分子层面解读中药方剂配伍机制提供了新的技术手段与方法学借鉴。

关键词

中药成分 / 分子相互作用 / 深度学习 / 跨域泛化 / 分子结构 / 配伍机制

Abstract

收起

Objective Given the complexity of traditional Chinese medicine (TCM) formula constituents and their unclear interactions, a deep learning-based molecular interaction prediction model is constructed taking the prediction of interactions among constituents as the entry point, providing methodological support for identifying potential constituent interactions and subsequent studies on formula compatibility mechanisms. Methods We propose a molecule-molecule interaction prediction model based on molecular 3D structural features and Extended Connectivity Fingerprints, designated as 3D structure and extended connectivity fingerprints molecule-molecule interaction (3DFPMMI). The model employs the SchNet network to extract deep features from the 3D graph structure of molecules while simultaneously acquiring Extended Connectivity Fingerprints, and achieves effective fusion of multi-source features through a contrastive learning strategy. The model is trained on the DrugBank drug-drug interaction dataset, and cross&8209;domain generalization is implemented to realize the prediction of interactions among TCM constituents. Results The 3DFPMMI model achieves an accuracy of 0.921 and an area under the precision-recall curve of 0.972 on the DrugBank dataset, both outperforming other benchmark models. Ablation studies confirm that molecular 3D structural features, extended connectivity fingerprint features, and the contrastive learning-based feature fusion strategy are all core effective modules for improving model performance. When applied to predict component interactions in the classic TCM formulas Buyang Huanwu Decoction and Banxia Xiexin Decoction, some of the prediction results are highly consistent with the conclusions from existing experimental studies. Conclusion This study establishes a novel method for predicting interactions among TCM ingredients and preliminarily validates its application potential. This work offers new technical tools and methodological references for interpreting the compatibility mechanisms of TCM formulas at the molecular level.

Key words

traditional Chinese medicine components / molecular interaction / deep learning / cross-domain generalization / molecular structure / compatibility mechanism

引用本文

程宁, 曾湘祥, 宋学坤. 基于分子结构跨域泛化的中药成分相互作用预测研究. 中草药, 2026 , 57 (17) : 6867 -6877 . DOI: 10.7501/j.issn.0253-2670.2026.17.018

CHENG Ning, ZENG Xiangxiang, SONG Xuekun. Prediction of traditional Chinese medicine component interactions via cross-domain generalization of molecular structure[J]. Chinese Traditional and Herbal Drugs, 2026 , 57 (17) : 6867 -6877 . DOI: 10.7501/j.issn.0253-2670.2026.17.018

基金

收起

国家中医药管理局中医药创新团队及人才支持计划项目 (ZYYCXTD-D-202408); 国家自然科学基金项目 (61702164); 河南省卫生健康委国家中医临床研究基地科研专项课题 (2021JDZX2135)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

Gong X, Sucher N J. Stroke therapy in traditional Chinese medicine (TCM): Prospects for drug discovery and development [J]. Phytomedicine, 2002, 9(5): 478-484.
Zhao L, Zhang H, Li N, et al. Network pharmacology, a promising approach to reveal the pharmacology mechanism of Chinese medicine formula [J]. J Ethnopharmacol, 2023, 309: 116306.
Wu J S, Zhang F Q, Li Z Z, et al. Integration strategy of network pharmacology in traditional Chinese medicine: A narrative review [J]. J Tradit Chin Med, 2022, 42(3): 479-486.
Wang H M, Xu M, Guo Z T, et al. DTI-BGCGCN: A novel bipartite graph-based cluster graph convolutional network for drug-target interaction prediction in modern and traditional Chinese medicine [J]. Comput Biol Chem, 2025, 117: 108410.
Lu Y Q, Chen J, Fan N N, et al. Machine learning models for drug-drug interaction prediction from computational discovery to clinical application [J]. NPJ Digit Med, 2026, 9: 198.
Hashir Q, Asfand E Yar M, Ullah A, et al. Computational approaches for drug-drug interaction prediction: A systematic review of data sources, modeling strategies, and evaluation frameworks [J]. Front Pharmacol, 2026, 17: 1816394.
Ahmad M M, Ali Shah Z B, Nies H W. A systematic review of molecular structures, knowledge graphs, and cold-start scenario in drug-drug interaction prediction [J]. Comput Biol Med, 2025, 190: 110122.
Zhang X X, Lei X J. Predicting miRNA-drug interactions via dual-channel network based on TCN and BiLSTM [J]. Front Comput Sci, 2024, 19(5): 195905.
Wang H P, Meng X J, Zhang Y. Biomolecular interaction prediction: The era of AI [J]. Adv Sci, 2025, 12(38): e09501.
le Maire A, Bourguet W. What structural biology tells us about the mode of action and detection of toxicants [J]. Annu Rev Pharmacol Toxicol, 2025, 65: 529-546.
Li X C, Chen K, Yang J X, et al. TLDA: A transfer learning based dual-augmentation strategy for traditional Chinese medicine syndrome differentiation in rare disease [J]. Comput Biol Med, 2024, 169: 107808.
Ren Z H, Zeng X X, Lao Y Z, et al. Predicting rare drug-drug interaction events with dual-granular structure-adaptive and pair variational representation [J]. Nat Commun, 2025, 16: 3997.
Budhkar A, Song Q Q, Su J, et al. Demystifying the black box: A survey on explainable artificial intelligence (XAI) in bioinformatics [J]. Comput Struct Biotechnol J, 2025, 27: 346-359.
Pan D P, Guo Y L, Fan Y F, et al. Development and application of traditional Chinese medicine using AI machine learning and deep learning strategies [J]. Am J Chin Med, 2024, 52(3): 605-623.
Knox C, Wilson M, Klinger C M, et al. DrugBank 6.0: The DrugBank knowledgebase for 2024[J]. Nucleic Acids Res, 2024, 52(D1): D1265-D1275.
张仲景, 钱超尘, 郝万山. 伤寒论[M]. 北京: 人民卫生出版社, 2023: 28-188.
王清任. 医林改错[M]. 北京: 中国中医药出版社, 2023: 5-78.
中国药典[S]. 一部. 2025.
孙景存, 张丁丁, 王嘉欣, 等. UPLC-MS/MS法同时测定补阳还五汤中17种成分的含量[J]. 中成药, 2025, 47(12): 3903-3908.
王若玲, 霍静娴, 张辉兰, 等. 半夏泻心汤化学成分、药理作用的研究进展及质量标志物的预测[J]. 世界中医药, 2024, 19(5): 719-726.
Gao K, Liu L, Lei S, et al. HERB 2.0: An updated database integrating clinical and experimental evidence for traditional Chinese medicine [J]. Nucleic Acids Res, 2025, 53(D1): D1404-D1414.
Schütt K T, Kindermans P J, Felix H E S, et al. SchNet: A continuous-filter convolutional neural network for modeling quantum interactions [A] // Advances in Neural Information Processing Systems 30[C]. Long Beach: Curran Associates, Inc., 2017: 992-1002.
Popescu M-C, Balas V E, Perescu-Popescu L, et al. Multilayer perceptron and neural networks [J]. WSEAS Transactions Circuits Systems, 2009, 8(7): 579-588.
Naidu G, Zuva T, Sibanda E M. A review of evaluation metrics in machine learning algorithms [A] // Artificial Intelligence Application in Networks and Systems [M]. Cham: Springer International Publishing, 2023: 15-25.
Deng Y F, Xu X R, Qiu Y, et al. A multimodal deep learning framework for predicting drug-drug interaction events [J]. Bioinformatics, 2020, 36(15): 4316-4322.
Qiao J B, Guo X, Jin J R, et al. Taco-DDI: Accurate prediction of drug-drug interaction events using graph transformer-based architecture and dynamic co-attention matrices [J]. Neural Netw, 2025, 189: 107655.
Abbas K, Hao C, Yong X, et al. Graph neural network-based drug-drug interaction prediction [J]. Sci Rep, 2025, 15: 30340.
Ryu J Y, Kim H U, Lee S Y. Deep learning improves prediction of drug-drug and drug-food interactions [J]. Proc Natl Acad Sci USA, 2018, 115(18): E4304-E4311.
Yin Q J, Fan R, Cao X S, et al. DeepDrug: A general graph-based deep learning framework for drug-drug interactions and drug-target interactions prediction [J]. Quant Biol, 2023, 11(3): 260-274.
Li M L, Cao C, et al. MHAFR-DDI: A multimodal hierarchical attention fusion and relation-aware architecture for drug-drug interaction event prediction [J]. Brief Bioinform, 2026, 27: bbag077.
Oruh J, Viriri S, Adegun A. Long short-term memory recurrent neural network for automatic speech recognition [J]. IEEE Access, 2022, 10: 30069-30079.
Franceschi L, Donini M, Perrone V, et al. Hyperparameter optimization in machine learning [J]. FNT Machine Learning, 2025, 18(6): 1054-1201.
Zheng K, Zhang Z, Du C, et al. Ferulic acid enhances the chemical and biological properties of Astragali Radix: A stimulator for Danggui Buxue Tang, an ancient Chinese herbal decoction [J]. Planta Med, 2014, 80(2/03): 159-164.
Ren J W, Ding Y T, Li S Z, et al. Predicting the anti-inflammatory mechanism of Radix Astragali using network pharmacology and molecular docking [J]. Medicine, 2023, 102(35): e34945.
Rao Y J, Li J J, Qiao R F, et al. Tetramethylpyrazine and Astragaloside IV have synergistic effects against spinal cord injury-induced neuropathic pain via the OIP5-AS1/ miR-34a/Sirt1/NF-κB axis [J]. Int Immunopharmacol, 2023, 115: 109546.
黄文娟, 刘晶, 陈良, 等. 黄连-姜的不同比例配伍使用对有效成分含量的影响[J]. 新疆医科大学学报, 2018, 41(10): 1298-1301.
李洁, 任爱农, 赵妍. 高效液相色谱法同时测定干姜甘草配伍后6-姜辣素、甘草酸和甘草苷的含量变化[J]. 中南药学, 2016, 14(9): 994-997.
尚文斌, 郭超, 于希忠, 等. 人参皂苷Rb₁与小檗碱配伍对糖尿病小鼠糖脂代谢的影响[J]. 时珍国医国药, 2015, 26(3): 518-521.

2026年第57卷第17期

PDF下载

引用本文

BibTeX

文章信息

doi: 10.7501/j.issn.0253-2670.2026.17.018

接收时间：2026-03-25
首发时间：2026-10-29

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2026-03-25

基金

作者信息

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/zcy/CN/10.7501/j.issn.0253-2670.2026.17.018

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT