進階搜尋


下載電子全文  
系統識別號 U0026-0708201721595600
論文名稱(中文) 基於深度學習建置廣泛性的潛在藥物不良反應預測模型
論文名稱(英文) A Generalized Model for Predicting Potential Adverse Drug Reactions by Deep Neural Network
校院名稱 成功大學
系所名稱(中) 資訊工程學系
系所名稱(英) Institute of Computer Science and Information Engineering
學年度 105
學期 2
出版年 106
研究生(中文) 林佩儒
研究生(英文) Pei-Ju Lin
學號 P76041239
學位類別 碩士
語文別 英文
論文頁數 33頁
口試委員 指導教授-蔣榮先
口試委員-高雅慧
口試委員-鄭靜蘭
口試委員-郝沛毅
口試委員-胡敏君
中文關鍵字 文獻探勘  藥物不良反應  深度學習  資訊擷取 
英文關鍵字 text mining  adverse drug reaction  deep learning  information retrieval 
學科別分類
中文摘要 根據世界衛生組織定義,藥物不良反應指藥物在人體上所產生一種不舒服,有害性或未預期的反應。藥物不良反應是用藥安全議題中一個極為重要且不容忽視的問題,特別是許多藥物由於在上市前有限的生物實驗和臨床試驗,因此許多潛在不良反應需仰賴著自發性的通報,但礙於自發性通報的通報率低且無法確認回報的真實性。因此若能結合更多有關藥物和不良反應的資訊,針對既有藥物判斷是否有其他潛在不良反應,將能使藥物安全監測更完善。
本研究目的是建立一個廣泛性的藥物不良反應預測模型,不論是否有此藥物的不良反應紀錄,本模型都能利用現有的資訊去預測潛在的不良反應。由於先前的研究大多將偵測潛在藥物不良反應視為一個分類問題,未考慮到藥物不良反應的預測是有時間次序性的,如果將該藥所有不良反應紀錄視為同一時間的特徵將會失去預測的時效性。此外,隨著藥物研發流程與實驗,藥物不良反應會在不同階段被偵測出來,相對也會有許多文獻資料發布,若能利用這些資料及特性,將可以呈現更完整的藥物特性於不良反應預測上。本模型利用資訊擷取的方法,整合藥物化學和生物特徵、生醫文獻特徵、藥物的表現型特徵,並透過深度學習來評估藥物與藥物不良反應的關係。
在實驗中,我們利用兩個年份的藥物與不良反應關係的資料模擬藥物在不同時間下不良反應紀錄的差異。在實驗結果說明我們提出的模型相較其他現有模型,有更好的平均準確率。在這項研究中,我們結合了藥物的生物化學特徵與文獻探勘技術預測更多可能的藥物不良反應,其中透過文獻探勘技術取得的特徵能有效提升系統預測的結果,且透過藥物敘述擴張模型可以從現有文獻紀錄中取得新藥物的特徵。最重要的是,不論是否有該藥物的不良反應紀錄,都可透過本模型預測出可能的潛在不良反應。
英文摘要 As defined by the World Health Organization, adverse drug reaction (ADR) is “a response to a medicine which is noxious and unintended, and which occurs at doses normally used in man . . . .” Adverse drug reaction is a crucial topic in drug safety that should not be neglected. Because of the limitation of clinical trials, the detection of ADRs relies on spontaneous reports. However, low reporting rates and the under-reporting of spontaneous report become a problem on pharmacovigilance. Therefore, if we can detect the adverse drug reaction earlier, the drug safety evaluators can assess the potential adverse drug reaction and enhance the pharmacovigilance.
Previous studies usually treat the prediction of adverse drug reaction as a classification problem without consideration of time. It ignores the chronological orders of the data of adverse drug reactions. During the drug development process, adverse drug reaction will be detected at different stages corresponding to a lot of published literature. These data are comprehensive properties of drugs that are used to predict the adverse drug reactions. Without the proper time order, the prediction of ADR will not be as accurate.
This study aims to develop a generalized adverse drug prediction model that can utilize the properties of the drug to predict potential ADR even without previous adverse drug reaction record. We used deep learning combining with chemical properties, biological properties and the drug embedding we extracted from literature mining to assess the potential adverse drug reaction. The drug embedding we learned from large-scale literature can effectively enhance the performance. The drug2vec expansion model can represent the drug we don’t have any property before. Most important of all, we can predict the potential adverse drug reaction whether or not the adverse drug reaction was recorded before.
論文目次 中文摘要 I
Abstract III
誌謝 V
Contents VI
List of Tables VIII
List of Figures IX
Chapter 1 Introduction 1
1.1 Background 1
1.2 Motivation 2
1.3 Research Objective 2
1.4 Thesis Organization 3
Chapter 2 Related Work 4
2.1 Pharmacovigilance 4
2.2 Detection of Adverse Drug Reaction 5
2.2.1 Detect ADR from Omics Data 5
2.2.2 Detect ADR from Biomedical Literature 6
Chapter 3 Materials and Methods 7
3.1 Study of Our Research Work 8
3.2 Data Collection 10
3.2.1 Collect the Drugs and the ADR 10
3.2.2 Extract Chemical and Biological Properties 11
3.3 Data Processing 13
3.3.1 ADR Vector Notation 13
3.3.2 Drug Embedding 14
3.3.3 Drug2Vec Expansion 15
3.4 Generalized Adverse Drug Reaction Prediction Model 16
3.4.1 Model Architecture 17
3.4.2 Model Building 18
Chapter 4 Experiments 20
4.1 Experimental Design 20
4.2 Survey of Drug Embedding 21
4.3 Performance Evaluation 22
4.3.1 Evaluation of Model 23
4.3.2 Evaluation of Features 24
4.3.3 Discussion on Model Depth 25
4.4 Evaluation of Generalized Model 26
4.4.1 Evaluation of Detection and Prediction 26
4.4.2 Evaluation of Drug2Vec Expansion 27
4.4.3 Case Study 27
Chapter 5 Conclusions and Future Work 29
5.1 Conclusions 29
5.2 Future Work 30
References 32
參考文獻 [1] WHO, "Safety of Medicines A guide to detecting and reporting adverse drug reactions," 2002.
[2] J. Lazarou, B. H. Pomeranz, and P. N. Corey, "Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies," Jama, vol. 279, no. 15, pp. 1200-1205, 1998.
[3] FDA. (2009). ADRs: Prevalence and incidence. Available: https://www.fda.gov/drugs/developmentapprovalprocess/developmentresources/druginteractionslabeling/ucm110632.htm#ADRs:%20Prevalence%20and%20Incidence
[4] T. D. R. Foundation, "Drug Safety Newsletter," 2017, vol. 57th.
[5] L. Hazell and S. A. Shakir, "Under-reporting of adverse drug reactions," Drug safety, vol. 29, no. 5, pp. 385-396, 2006.
[6] J. S. Brown et al., "Early detection of adverse drug events within population‐based health networks: application of sequential testing methods," Pharmacoepidemiology and drug safety, vol. 16, no. 12, pp. 1275-1284, 2007.
[7] J. S. Almenoff, K. K. LaCroix, N. A. Yuen, D. Fram, and W. DuMouchel, "Comparative performance of two quantitative safety signalling methods," Drug safety, vol. 29, no. 10, pp. 875-887, 2006.
[8] B. Hughes, "2007: spotlight on drug safety: what have been some of the past year's key events and trends for the biopharma industry?," Nature Reviews Drug Discovery, vol. 7, no. 1, pp. 5-8, 2008.
[9] A. Cami, A. Arnold, S. Manzi, and B. Reis, "Predicting Adverse Drug Events Using Pharmacological Network Models," Science Translational Medicine, vol. 3, no. 114, pp. 114ra127-114ra127, 2011.
[10] W. H. Organization, "The importance of pharmacovigilance," 2002.
[11] T. D. R. Foundation, "Drug Injury Relief Act," 2014, vol. 13th.
[12] S. Karimi, C. Wang, A. Metke-Jimenez, R. Gaire, and C. Paris, "Text and Data Mining Techniques in Adverse Drug Reaction Detection," ACM Comput. Surv., vol. 47, no. 4, pp. 1-39, 2015.
[13] T. Lorberbaum, M. Nasir, M. J. Keiser, S. Vilar, G. Hripcsak, and N. P. Tatonetti, "Systems pharmacology augments drug safety surveillance," Clinical Pharmacology & Therapeutics, vol. 97, no. 2, pp. 151-158, 2015.
[14] L.-C. Huang, X. Wu, and J. Y. Chen, "Predicting adverse drug reaction profiles by integrating protein interaction networks with drug structures," PROTEOMICS, vol. 13, no. 2, pp. 313-324, 2013.
[15] M. Liu et al., "Large-scale prediction of adverse drug reactions using chemical, biological, and phenotypic properties of drugs," Journal of the American Medical Informatics Association, vol. 19, no. e1, pp. e28-e35, 2012.
[16] J. Finkelstein, Q. Chen, H. Adams, and C. Friedman, "Automated Summarization of Publications Associated with Adverse Drug Reactions from PubMed," AMIA Summits on Translational Science Proceedings, vol. 2016, p. 68, 2016.
[17] W. Wang, K. Haerian, H. Salmasian, R. Harpaz, H. Chase, and C. Friedman, "A drug-adverse event extraction algorithm to support pharmacovigilance knowledge mining from PubMed citations," in AMIA Annu Symp Proc, 2011, vol. 2011, pp. 1464-1470.
[18] H. Gurulingappa, A. Mateen‐Rajpu, and L. Toldo, "Extraction of potential adverse drug events from medical case reports," Journal of biomedical semantics, vol. 3, no. 1, p. 15, 2012.
[19] N. Arthur et al., "The Importance of Pharmacovigilance-Safety Monitoring of Medicinal Products," World Health Organization, 2002.
[20] M. Kuhn, I. Letunic, L. J. Jensen, and P. Bork, "The SIDER database of drugs and side effects," Nucleic acids research, p. gkv1075, 2015.
[21] E. E. Bolton, Y. Wang, P. A. Thiessen, and S. H. Bryant, "PubChem: integrated platform of small molecules and biological activities," Annual reports in computational chemistry, vol. 4, pp. 217-241, 2008.
[22] D. S. Wishart et al., "DrugBank: a comprehensive resource for in silico drug discovery and exploration," Nucleic acids research, vol. 34, no. suppl 1, pp. D668-D672, 2006.
[23] T. Mikolov, K. Chen, G. Corrado, and J. Dean, "Efficient estimation of word representations in vector space," arXiv preprint arXiv:1301.3781, 2013.
[24] R. Kiros et al., "Skip-thought vectors," in Advances in neural information processing systems, 2015, pp. 3294-3302.
[25] R. Tibshirani, "Regression shrinkage and selection via the lasso," Journal of the Royal Statistical Society. Series B (Methodological), pp. 267-288, 1996.
論文全文使用權限
  • 同意授權校內瀏覽/列印電子全文服務,於2018-08-08起公開。
  • 同意授權校外瀏覽/列印電子全文服務,於2018-08-08起公開。


  • 如您有疑問,請聯絡圖書館
    聯絡電話:(06)2757575#65773
    聯絡E-mail:etds@email.ncku.edu.tw