臺灣博碩士論文加值系統

肺癌是全球主要死亡率最高的疾病之一，其相關研究已是近來熱門的研究主題。藥物重新定位是一種有效率的策略，對於不同疾病之現有藥物找出其是否可能應用於治療其他疾病的過程。藥物重新定位提供了舊藥新用和避免副作用的最佳平衡性，此方法已有一些成功的案例；而使用深度學習於巨量資料中過濾與學習隱藏的規則發掘有用的資訊以解決問題也已是一種趨勢。本論文中，我們提出利用藥物重新定位策略整合深度學習策略與突變基因資料以辨識治療非小細胞肺癌的潛在藥物。
本研究整合兩項資訊：癌症蛋白預測與突變基因方法以發展一套新的藥物重新定位流程，由多個層面，如蛋白質互動資料、作用域資料庫、GO ID、TCGA及癌症關聯基因等，分析肺癌資料集。應用深度學習方法與5種特徵資料集，我們建立了一個高信賴度的癌症蛋白分類法，此分類法的預測效能F1 最高可達82%。本研究整合癌症蛋白預測與突變基因方法取得關鍵基因集，並且依據此關鍵基因集合從cMap基因藥物資料庫中推導潛在的標靶治療藥物。
本研究透過文獻探討、臨床試驗與IC50實驗等方式驗證標靶藥物的可信度。透過整合深度學習策略與突變基因方法，本論文所預測之非小細胞肺癌標靶藥物的IC50實驗準確度可達53%，較以往之方法來得高，由此可證明此研究所提出之非小細胞肺癌的藥物重新定位流程的效能。

Lung Cancer is one of the diseases that leading causes of death worldwide and it is a hot research topic. Drug repositioning is an effective approach for identifying and developing potential new therapeutic opportunities for existing drugs in a different disease. Drug repositioning provides the best tradeoff between drug discovery and risk has yielded some successes. On the other hand, mining the hidden rules and deriving useful information by deep learning technique has been a trend. In this thesis, we propose to make use of the drug repositioning approach to identify potential drugs for non-small cell lung cancer by integrating deep learning approach and gene mutation data.
This work integrates two pieces of information - cancer protein prediction and gene mutation data – to develop a novel pipeline of drug repositioning to analyze lung cancer datasets from protein-protein interaction, domain database, GO ID, TCGA, and cancer-associate genes. High confident cancer protein prediction is obtained by well-training a deep learning approach using five molecular features. The proposed approach can achieve an F1 ratio of 82% for lung cancer protein prediction. In this study, the sets of key genes are first derived from deep learning approach and gene mutation data, and then they are used to infer potential-repositioning drugs from the cMap database.
The efficiency of these drugs is supported from the literature review, experimentally determined in-vitro IC50 and clinical trials. This work provides better drug prediction accuracy (53%) than some competing methods, which demonstrates the efficiency of the proposed drug repositioning pipeline.

Chinese Abstract……………………………………………………….…………...........i
English Abstract……………………………………………………….………….........iii
Acknowledgement……………………………………………………….………..........v
Table of Contents……………………………………………………….….……..........vi
List of Tables……………………………………………………….…….……..............vii
List of Figures……………………………………………………….…….…….............viii
List of Abbreviations……………………………………………………….…….……...ix
Chapter 1 Introduction……………………………………………..………..............1
1.1 Research Motivation……………………………………………......................1
1.2 Thesis Objectives……………………………………………….......................2
1.3 Thesis Outline…………………………………………………..........................3
Chapter 2 Background and Related Works……………………...…………....4
2.1 Lung cancer……………………………………………………...........................4
2.2 Drug repositioning…………………………………………….........................4
2.3 Machine learning……………………………………………….........................5
2.4 The relationship between gene mutation and cancer…………….....5
2.5 Related Work................................................................................6
Chapter 3 Methodology……………………………………………………...............8
3.1 System Flowchart………………………………………………........................8
3.2 Input datasets…………………………………………………….......................10
3.3 Identifications of DEGs………………………………………….....................11
3.4 Feature Score Generation………………………………………...................12
3.4.1 Domain-domain Interaction (DDI)……………………………................12
3.4.2 Weighted Domain Frequency Score (DFS)……………………..........13
3.4.3 Cancer Linker Degree (CLD)…………………………………..................13
3.4.4 FFI feature……………………………………………………..........................14
3.5 Data Normalization……………………………………………........................14
3.6 Deep learning method…………………………………………......................15
3.6.1 H2O -deep learning package……………………………………...............17
3.6.2 Grid search method…………………………………………….....................18
3.6.3 Cross-validation………………………………………………….....................18
3.7 Gene mutation data……………………………………………........................19
3.8 Drug discovery…………………………………………………..........................19
3.9 Perform Robustness Analysis…………………………….........................20
3.10 Comparison of the procedure of three methods.........................21
Chapter 4 Results and Discussions………………………………………….........25
4.1 Results of data preprocessing…………………………………...................25
4.2 Results of deep learning method……………………………….................25
4.3 Results of gene mutation method ………………………………...............27
4.4 Drug discovery for treating NSCLC……………………………................28
4.5 Retrieving target genes for lung cancer…………………….................30
4.6 Target genes-related PPI………………………………………....................31
4.7 Results of the robustness analysis……………………………….............33
4.8 Comparison with competing research………………………….............36
Chapter 5 Conclusion and Future work………………………………….….....38
Reference…………………………………………………….……………....................39
Appendix…………………………………………………………………......................42
Extended Abstract………………………………………………………………….......57

[1] Berman AT, James SS, Rengan R. Structure mechanism, and evolution of the mRNA capping apparatus. Cancers (Basel). 2015; 7(3): 1178-90.
[2] Shim JS, Liu JO. Recent Advances in Drug Repositioning for the Discovery of New Anticancer drugs. Int. J. Biol. Sci. 2014; 10(7): 654-663.
[3] Trinh LD, Michael P, Michael AW, Gunda IG. Chapter 23 – The Future of Drug Repositioning: Old Drugs, New Opportunities. Annual Reports in Medicinal Chemistry. 2011; 46: 385-401.
[4] Stachnik A, Yuen T, Iqbal J, Sgobba M, Gupta Y, Lu P, et al. Repurposing of bisphosphonates for the prevention and therapy of nonsmall cell lung and breast cancer. Proc Natl Acad Sci U S A. 2014; 111(50): 17995–8000.
[5] Chen H, Zhang H, Zhang Z, Cao Y, Tang W. Network-based inference methods for drug repositioning. Comput Math Methods Med. 2015; 2015: 130620. 

[6] Lee HS, Bae T, Lee JH, Kim DG, Oh YS, Jang Y, et al. Rational drug repositioning guided by an integrated pharmacological network of protein, disease and drug. BMC Syst Biol. 2012; 6: 80.
[7] Huang CH, Peng HS, and Ng KL. Prediction of Cancer Proteins by Integrating Protein Interaction, Domain Frequency, and Domain Interaction Data Using Machine Learning Algorithms. BioMed Res Int. 2015; 2015: 312047.
[8] Nordling CO. A new theory on cancer-inducing mechanism. Br J Cancer. 1953; 7(1): 68–72. 

[9] Armitage P, Doll R. The age distribution of cancer and a multi-stage theory of carcinogenesis. Br J Cancer. 1954; 8(1): 1–12.
[10] Cristian T, Luigi M, Martin A. Nowak, Giovanni Parmigianni, and Bert Vogelstein. Only three driver gene mutations are required for the development of lung and colorectal cancers. PNAS. 2015; 112(1): 118-123.
[11] Bert V, Nickolas P, Victor EV, Shibin Z, Luis AD and Kenneth WK. Cancer Genome Landscapes. Science. 2013 March 29; 339(6127): 1546-1558.
[12] Hurle MR. Yang L, Xie Q, Rajpal DK, Sanseau P, Agaewal P. Computational drug repositioning: from data to therapeutics. Clin Pharmacol Ther. 2013; 93(4): 335-41.
[13] Alexander A, Sergey P, Artem A, Alvaro U, Polina M, Alex Z. Deep learning applications for predicting pharmacological properties of drugs and drug repurposing using transcriptomic data. Mol Pharm. 2016; 13(7): 2524-2530.
[14] Padideh D, Reza G. A Deep Learning Approach for Cancer Detection and Relevant Gene Identification. Pacific Symposium on Biocomputing. 2017.
[15] Paul B, Berend O, Sam S. Deep Learning and identification of Cancer Related Sub-networks. Bachelor Project. 2016.
[16] Lyu B, Haque A. Deep Learning Based Tumor Type Classification Using Gene Expression Data. bioRxiv. 2018.
[17] Yuan Y, Shi Y, Li C, and et al. DeepGene: an advanced cancer type classifier based on deep learning and somatic point mutations. BMC Bioinformatics. 2016; 17: 476.
[18] Wen M, Zhang Z, Niu S, Sha H, Yang R, Yun Y, Lu H. Deep-Learning-Based Drug-Target Interaction Prediction. J. Proteome Res. 2017; 1401-1409.
[19] Huang CH, Chang PM, Hsu CW, Huang CY, and Ng KL. Drug repositioning for non-small cell lung cancer by using machine learning algorithms and topological graph theory. Huang et al. BMC Bioinformatics. 2016; 17 Suppl 1:2.
[20] Zhao XM, Chen L. Domain-Domain Interaction Identification with a Feature Selection Approach. PRIB. 2008; 178-186.
[21] Chan H. Identification of novel tumor-associated gene (TAG) by bioinformatics analysis [M.S. thesis]. National Cheng Kung University, Tainan City, Taiwan. 2006.
[22] Aragues R, Sander C, and Oliva B. “Predicting cancer involvement of genes from heterogeneous data,” BMC Bioinformatics. vol. 9, article 172, 2008.
[23] Kochura Y, Stirenko S, Gordienko Y. Comparative Performance Analysis of Neural Networks Architectures on H2O Platform for Various Activation Functions. IEEE International Young Scientists Forum on Applied Physics and Engineering. 2017.
[24] Yenumula BR. Classification and automatic recognition of objects using H2O package. SPIE: 10185. 2017.
[25] James B, Yoshua B. Random Search for Hyper-Parameter Optimization. Journal of Machine Learning Research. 2012; 281-305.
[25] Payam R, Lei T, Huan L. Cross-Validation. B Vijayalakshmi Galleys. 2008.
[27] Wolf FM. Meta-Analysis: Quantitative Methods for Research Synthesis. California: Sage publications. 1986.
[28] Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to meta-analysis. United Kingdom: Wiley press. 2009.
[29] Alhussein F, Seyed-Moshen MD, Pascal F, A Geometric Perspective on the Robustness of Deep Networks. IEEE Signal Processing Magazine. 2017; 34: 6.
[30] Zheng S, Song Y, Leung T, Goodfellow Ian. Improving the Robustness of Deep Neural Networks via Stability Training. IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016; 4480-4488.
[31] Schott L, Rauber J, Bethge M, Brendel W. Towards the First Adversarially Robust Neural Network Model on MNIST. arXiv. 2018.
[32] Breitkreutz BJ, Stark C, Reguly T, Boucher L, Breitkreutz A, Livstone M, et al. The BioGRID Interaction Database: 2008 update. Nucleic Acids Res. 2008; 36(Database issue): D637–40.
[33] Huang CH, Chang PM, Lin YJ, Wang CH, Huang CY, and Ng KL. Drug Repositioning Discovery for Early- and Late-Stage Non-Small-Cell Lung Cancer. BioMed Res Int. 2014; 2014: 193817.
[34] Huang CH, Wu MY, Chang PM, Huang CY, Ng KL. In silico identification of potential targets and drugs for non-small cell lung cancer. IET Systems Biology. 2014; 8(2), p. 56-66.
[35] Yeh CT, Wu TH, Chang PM, and et al. Trifluoperazine, an Antipsychotic Agent, Inhibits Cancer Stem Cell Growth and Overcomes Drug Resistance of Lung Cancer. Am J Respir Crit Care Med. 2012; 186(11): 1180-1188.
[36] Wishart DS, Knox C, Guo AC, Cheng D, Shrivastava S, Tzur D, et al. DrugBank: a knowledgebase for drugs, drug actions, and drug targets. Nucleic Acids Res. 2008; 36(Database issue): D901–6.