广东工业大学学报 ›› 2024, Vol. 41 ›› Issue (01): 19-26,40.doi: 10.12052/gdutxb.230031

• 智慧医疗 • 上一篇    下一篇

基于知识图谱嵌入与深度学习的药物不良反应预测

吴菊华, 李俊锋, 陶雷   

  1. 广东工业大学 管理学院, 广东 广州 510520
  • 收稿日期:2023-02-22 出版日期:2024-01-25 发布日期:2024-02-01
  • 通信作者: 李俊锋(1995–) ,男,硕士研究生,主要研究方向为数据挖掘、机器学习,E-mail:751947938@qq.com
  • 作者简介:吴菊华(1974–) ,女,教授,主要研究方向为智慧健康管理、商务智能和工业互联网等
  • 基金资助:
    国家自然科学基金资助面上项目(71771059);广东省基础与应用基础研究基金资助项目(2021A1515220031)

Prediction of Adverse Drug Reactions Based on Knowledge Graph Embedding and Deep Learning

Wu Ju-hua, Li Jun-feng, Tao Lei   

  1. School of Management, Guangdong University of Technology, Guangzhou 510520, China
  • Received:2023-02-22 Online:2024-01-25 Published:2024-02-01

摘要: 识别药物潜在的不良反应,有助于辅助医生进行临床用药决策。针对以往研究的特征高维稀疏、需要为每种不良反应构建独立预测模型且预测精度较低的问题,本文开发一种基于知识图谱嵌入和深度学习的药物不良反应预测模型,能够对实验所覆盖的不良反应进行统一预测。一方面,知识图谱及其嵌入技术能够融合药物之间的关联信息,缓解特征矩阵高维稀疏的不足;另一方面,深度学习的高效训练能力能够提升模型的预测精度。本文使用药物特征数据构建药物不良反应知识图谱;通过分析不同嵌入策略下知识图谱的嵌入效果,选择最佳嵌入策略以获得样本向量;然后构建卷积神经网络模型对不良反应进行预测。结果表明,在DistMult嵌入模型和400维嵌入策略下,卷积神经网络模型预测效果最佳;重复实验的准确率、F1分数、召回率和曲线下面积的平均值分别为0.887、0.890、0.913和0.957,优于文献报道中的方法。所得预测模型具有较好的预测精度和稳定性,可以为安全用药提供有效参考。

关键词: 药物不良反应, 知识图谱嵌入, 深度学习, 预测模型

Abstract: Identifying potential adverse reactions of drugs can help doctors make clinical medication decisions. In view of the high-dimensional sparse features of previous studies and low prediction accuracy in constructing an independent prediction model for each adverse reaction, a prediction model of adverse reactions based on knowledge graph embedding and deep learning is developed, which can uniformly predict the adverse reactions covered by the experiment. On the one hand, knowledge graph and its embedding technology can fuse the correlation information between drugs and alleviate the deficiency of high-dimensional sparse feature matrix. On the other hand, the efficient training ability of deep learning can improve the prediction accuracy. In the study, drug characteristic data is used to construct a knowledge graph of adverse drug reactions; by analyzing the embedding effect of different embedding strategies, the best embedding strategy is selected to obtain the sample vector. Then a convolutional neural networks model is constructed to predict adverse reactions. The results show that the convolutional neural networks model has the best prediction effect under the DistMult embedding model and the 400-dimensional embedding strategy. The mean values of accuracy, F1 score, recall and Area Under Curve were 0.887, 0.890, 0.913 and 0.957, respectively, which are better than those reported in the literature. The prediction model has good prediction accuracy and stability, which can provide an effective reference for safe medication.

Key words: adverse drug reaction, knowledge graph embedding, deep learning, prediction model

中图分类号: 

  • TP399
[1] WESTER K, JONSSON A K, SPIGSET O, et al. Incidence of fatal adverse drug reactions: a population based study [J]. British Journal of Clinical Pharmacology, 2008, 65(4): 573-579.
[2] COCOS A, FIKS A G, MASINO A J. Deep learning for pharmacovigilance: recurrent neural network architectures for labeling adverse drug reactions in Twitter posts [J]. Journal of the American Medical Informatics Association, 2017, 24(4): 813-821.
[3] JONATHAN H W, TERRY M, JAN D H. Cost of prescription drug-related morbidity and mortality [J]. Annals of Pharmacotherapy, 2018, 52(9): 829-837.
[4] 朱笑笑, 杨尊琦, 刘婧. 基于Bi-LSTM和CRF的药品不良反应抽取模型构建[J]. 数据分析与知识发现, 2019, 3(2): 90-97.
ZHU X X, YANG Z Q, LIU J. Construction of an adverse drug reaction extraction model based on Bi-LSTM and CRF [J]. Data Analysis and Knowledge Discovery, 2019, 3(2): 90-97.
[5] 国家药品不良反应监测年度报告(2018年) [J]. 中国药物评价, 2019, 36(6) : 476-480
[6] 郭凯. 基于深度学习和语义分析的药物不良反应发现[D]. 大连: 大连理工大学, 2017.
[7] 朱嘉静. 基于机器学习的药物不良反应关键问题研究[D]. 成都: 电子科技大学, 2020.
[8] LEE C Y, CHEN Y P. Machine learning on adverse drug reactions for pharmacovigilance [J]. Drug Discovery Today, 2019, 24(7): 1332-1343.
[9] LEE C Y, CHEN Y P. Prediction of drug adverse events using deep learning in pharmaceutical discovery [J]. Briefings in Bioinformatics, 2020, 22(2): 1884-1901.
[10] World Health Organization. International drug monitoring, the role of national centres: report of WHO Technical Group[R]. Geneva: WHO, 1972.
[11] RING J, BROCKOW K. Adverse drug reactions: mechanisms and assessment [J]. European Surgical Research, 2002, 34(1-2): 170-175.
[12] 谈志远, 赵荣生. 人工智能技术在药物不良反应监测与上报中应用的研究进展[J]. 临床药物治疗杂志, 2019, 17(2): 23-27.
TAN Z Y, ZHAO R S. Progress of studies of artificial intelligence in surveillance and report of adverse drug reactions [J]. Clinical Medication Journal, 2019, 17(2): 23-27.
[13] 赵霞, 陈瑶, 廖俊, 等. 基于医药大数据的药品不良反应信号挖掘探讨[J]. 中华医院管理杂志, 2017, 33(5): 4.
ZHAO X, CHEN Y, LIAO J, et al. Signal mining for adverse drug reactions based on healthcare big data: methodology and applications [J]. Chinese Journal of Hospital Administration, 2017, 33(5): 4.
[14] VOSS E A, BOYCE RD, RYAN P B, et al. Accuracy of an automated knowledge base for identifying drug adverse reactions [J]. Journal of Biomedical Informatics, 2017, 66: 72-81.
[15] 陈瑶, 吴红, 葛卫红, 等. 基于深度学习模型的我国药品不良反应报告实体关系抽取研究[J]. 中国药科大学学报, 2019, 50(6): 753-759.
CHEN Y, WU H, GE W H, et al. Research on entity relation extraction of Chinese adverse drug reaction reports based on deep learning method [J]. Journal of China Pharmaceutical University, 2019, 50(6): 753-759.
[16] 申晨, 林鸿飞. 基于图嵌入的社交媒体药物不良反应事件检测方法[J]. 大连理工大学学报, 2020, 60(5): 547-554.
SHEN C, LIN H F. Detection method of adverse drug events from social media based on graph embedding [J]. Journal of Dalian University of Technology, 2020, 60(5): 547-554.
[17] 仲雨乐, 马诗雯, 陆豪杰, 等. 基于机器学习的药品不良反应实体识别研究综述[J]. 软件工程, 2022, 25(8): 1-6.
ZHONG Y L, MA S W, LU H J, et al. Survey of research on entity recognition of adverse drug reaction based on machine learning [J]. Software Engineering, 2022, 25(8): 1-6.
[18] LIU M, WU Y H, CHEN Y K, et al. Large-scale prediction of adverse drug reactions using chemical, biological, and phenotypic properties of drugs [J]. Journal of the American Medical Informatics Association, 2012, 19(e1): e28-35.
[19] VILAR S, HRIPCSAK G. The role of drug profiles as similarity metrics: applications to repurposing, adverse effects detection and drug-drug interactions [J]. Briefings in Bioinformatics, 2017, 18(4): 670-681.
[20] MUNOZ E, NOVACEK V, VANDENBUSSCHE P Y. Using drug similarities for discovery of possible adverse reactions[J]. Amia Annual Symposium Proceedings, 2016, 2016: 924-933
[21] WANG C S, LIN P J, CHENG C L, et al. Detecting potential adverse drug reactions using a deep neural network model [J]. Journal of Medical Internet Research, 2019, 21(2): e11016.
[22] DAI Y F, WANG S P, NEAL N, et al. A survey on knowledge graph embedding: approaches, applications and benchmarks [J]. Electronics, 2020, 9(5): 750-778.
[23] ZENG X X, TU X Q, LIU Y S, et al. Toward better drug discovery with knowledge graph[J]. Current Opinion in Structural Biology, 2022, 72: 114-126
[24] SHTAR G, ROKACH L, SHAPIRA B. Detecting drug-drug interactions using artificial neural networks and classic graph similarity measures [J]. PloS One, 2019, 14(8): e0219796.
[25] JOSHI P, MASILAMANI V, MUKHERJEE A. A knowledge graph embedding based approach to predict the adverse drug reactions using a deep neural network [J]. Journal of Biomedical Informatics, 2022, 132: 104122-104131.
[26] ZHANG F, SUN B, DIAO X L, et al. Prediction of adverse drug reactions based on knowledge graph embedding [J]. BMC Medical Informatics and Decision Making, 2021, 21(1): 38-48.
[27] WANG M, MA X Y, SI J W, et al. Adverse drug reaction discovery using a tumor-biomarker knowledge graph[J]. Frontiers in Genetics, 2021, 11: 625659
[28] DEY S, LUO H, FOKOUE A, et al. Predicting adverse drug reactions through interpretable deep learning framework [J]. BMC Bioinformatics, 2018, 19(Suppl 21): 476-488.
[29] BEAN D M, WU H H, IQBAL E, et al. Knowledge graph prediction of unknown adverse drug reactions and validation in electronic health records [J]. Scientific Reports, 2017, 7(1): 16416.
[30] WISHART D S, FEUNANG Y D, GUO A C, et al. DrugBank 5.0: a major update to the DrugBank database for 2018 [J]. Nucleic Acids Research, 2018, 46(D1): D1074-D1082.
[31] KUHN M, LETUNIC I, JENSEN L J, et al. The SIDER database of drugs and side effects [J]. Nucleic Acids Research, 2016, 44(D1): D1075-D1084.
[32] YANG B S, YIH W T, HE X D, et al. Embedding entities and relations for learning and inference in knowledge bases[EB/OL]. arXiv: 1412.6575 (2015-08-29) [2023-02-22]. https://doi.org/10.48550/arXiv.1412.6575.
[33] NICKEL M, LORENZO R, TOMASO P. Holographic embeddings of knowledge graphs[C] //Thirtieth AAAI Conference on Artificial Intelligence. California: AAAI. 2016: 1955-1961.
[34] 高飞萌, 宋智慧. 艾司奥美拉唑镁肠溶片致肾损伤[J]. 药物不良反应杂志, 2020, 22(6): 381-382.
GAO F M, SONG Z H. Kidney injury caused by esomeprazole magnesium enteric-coated tablets [J]. Adverse Drug Reactions Journal, 2020, 22(6): 381-382.
[35] DHAKRIT R, IVES A L, HARRIOTT N G, et al. Comparative incidence of acute kidney injury in patients on vancomycin therapy in combination with cefepime, piperacillin-tazobactam or meropenem [J]. Journal of Chemotherapy (Florence, Italy) , 2021, 34(2): 1-7.
[36] NIITSU T, HAYASHI T, UCHIDA J, et al. Drug-induced kidney injury caused by osimertinib: report of a rare case [J]. Nephron, 2021, 146(1): 58-63.
[37] WAKABAYASHI K, YAMAMOTO S, HARA S, et al. Nivolumab-induced membranous nephropathy in a patient with stage IV lung adenocarcinoma [J]. CEN Case Rep, 2022, 11(2): 171-176.
[38] 苏伟, 苏培. 碘海醇注射液致患者获得性急性肾功能衰竭的不良反应[J]. 中国生化药物杂志, 2016, 36(10): 92-94.
SU W, SU P. Effect of adverse reaction of acquired acute renal failure caused by iohexol injection [J]. Chinese Journal of Biochemical and Pharmaceuticals, 2016, 36(10): 92-94.
[1] 梁宇辰, 蔡念, 欧阳文生, 谢依颖, 王平. 基于切片关联信息的慢性阻塞性肺疾病CT诊断[J]. 广东工业大学学报, 2024, 41(01): 27-33.
[2] 温雯, 刘莹, 蔡瑞初, 郝志峰. 面向多粒度交通流预测的时空深度回归模型[J]. 广东工业大学学报, 2023, 40(04): 1-8.
[3] 钟耿君, 李东. 基于通道分离机制的双分支点云处理网络[J]. 广东工业大学学报, 2023, 40(04): 18-23.
[4] 金宇凯, 李志生, 欧耀春, 张华刚, 曾江毅, 陈搏超. 基于多阶段聚类的PM2.5质量浓度预测及对比研究[J]. 广东工业大学学报, 2023, 40(03): 17-24.
[5] 吴俊贤, 何元烈. 基于通道注意力的自监督深度估计方法[J]. 广东工业大学学报, 2023, 40(02): 22-29.
[6] 刘冬宁, 王子奇, 曾艳姣, 文福燕, 王洋. 基于复合编码特征LSTM的基因甲基化位点预测方法[J]. 广东工业大学学报, 2023, 40(01): 1-9.
[7] 徐伟锋, 蔡述庭, 熊晓明. 基于深度特征的单目视觉惯导里程计[J]. 广东工业大学学报, 2023, 40(01): 56-60,76.
[8] 刘洪伟, 林伟振, 温展明, 陈燕君, 易闽琦. 基于MABM的消费者情感倾向识别模型——以电影评论为例[J]. 广东工业大学学报, 2022, 39(06): 1-9.
[9] 章云, 王晓东. 基于受限样本的深度学习综述与思考[J]. 广东工业大学学报, 2022, 39(05): 1-8.
[10] 郑佳碧, 杨振国, 刘文印. 基于细粒度混杂平衡的营销效果评估方法[J]. 广东工业大学学报, 2022, 39(02): 55-61.
[11] Gary Yen, 栗波, 谢胜利. 地球流体动力学模型恢复的长短期记忆网络渐进优化方法[J]. 广东工业大学学报, 2021, 38(06): 1-8.
[12] 赖峻, 刘震宇, 刘圣海. 基于全局数据混洗的小样本数据预测方法[J]. 广东工业大学学报, 2021, 38(03): 17-21.
[13] 岑仕杰, 何元烈, 陈小聪. 结合注意力与无监督深度学习的单目深度估计[J]. 广东工业大学学报, 2020, 37(04): 35-41.
[14] 曾碧, 任万灵, 陈云华. 基于CycleGAN的非配对人脸图片光照归一化方法[J]. 广东工业大学学报, 2018, 35(05): 11-19.
[15] 陈旭, 张军, 陈文伟, 李硕豪. 卷积网络深度学习算法与实例[J]. 广东工业大学学报, 2017, 34(06): 20-26.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!