氣喘是一種慢性炎症的呼吸道疾病。近期的多項研究證據指出氣喘與呼吸道微生物相(airway microbiome)及其功能有關。在本研究中，我們對孩童的呼吸道進行總基因體學分析(metagenomics analysis)來得到微生物豐富度的資訊，並且也分析其血清中的代謝體成分與濃度，最後將此兩類資料做綜合性的分析。為了瞭解呼吸道微生物相與循環代謝體如何影響氣喘，在此研究中探討三個研究目的。第一、釐清微生物相在氣喘病患與正常的人之間是否有明顯的差異，我們透過計算基因數量與探討呼吸道微生物多樣性來進行分析。第二、透過變數選取並搭配Random Forest與gradient boosting兩種模型，來找出與氣喘相關的呼吸道微生物與血清代謝體。並且利用Canonical Correspondence analysis 與Co-inertia analysis兩種方法來呈現被選取的重要微生物與代謝體之間的關聯性。最後，進行微生物功能層級的分析，根據上述兩個目標的結果來理解哪些微生物所參與的人體生理功能與氣喘相關。

我們發現多種微生物與代謝體被模型選出，並且其與氣喘的關聯已被研究且發表，例如：Prevotella spp., Neisseria spp., 麩醯胺酸(glutamine)與組胺酸(histidine)。多種微生物所參與的生理功能也在氣喘病患和正常人之間有明顯差異，特別的是醣胺聚醣降解(Glycosaminoglycan degradation)與組胺酸代謝(histidine metabolism)此兩類生理功能不僅與氣喘相關也與發炎反應相關。除此之外，Neisseria該屬之細菌參與該二個功能之占比，在兩組之間有明顯的差異。但是透過關聯性的分析，我們發現微生物與代謝體之間整體的關聯性並不強。總結來說，我們建立了一個完整的統計分析流程以分析人體呼吸道總基因體學之資料並使我們更進一步地理解氣喘、呼吸道微生物與血清代謝體三者之間的關係。

Asthma is a chronic inflammatory disease of the airways. Emerging evidence has related asthma to airway microbiome and their functions. In this study, we applied a metagenomics analysis on children’s airway. We also analyzed serum metabolites to do an integrated analysis with airway microbiome. In order to understand how airway microbiome and circulating metabolites affect childhood asthma, there are several goals in our study. First, we aimed to figure out the relationship of asthma with the alteration of airway microbiome between the case and control. We explored the difference in gene counts and biodiversity. Secondly, we identified the influential microorganisms and metabolites for asthma through feature selection with Random Forest and gradient boosting model. We also demonstrated the relationship between the selected microorganisms and metabolites with Canonical Correspondence analysis and Co-inertia analysis plot. Finally, we studied those microbial functions associated with asthma according to the findings in the first two aims.

We found that several selected microorganisms and metabolites are associated with asthma, and the association has been reported in the literature, such as Prevotella spp., Neisseria spp., glutamine and histidine. Many microbial functions are also found to be differentially enriched between the case and control. Specifically, two of those functions - Glycosaminoglycan degradation and histidine metabolism - are related to immune response, and several species in the genus Neisseria, which is known to be related to the two functions, have different abundance levels between normal and asthmatic samples. However, there is no overall significant difference of airway microbiome counts between the case and control and, the relationship between those important microbiome and metabolites is not strong. In conclusion, we built up a comprehensive statistical procedure to analyze metagenomics data from human airway and our finding provides a great understanding in the relationship among asthma, airway microbiome and serum metabolites.

1. Introduction 1
2. Materials and Methods 4
2.1. Gene count and rarefaction curve 4
2.2. Diversity 4
2.2.1. Alpha diversity 4
2.2.2. Beta diversity 7
2.3. Models for identifying asthmatic biomarkers 8
2.3.1. Random Forest (RF) 8
2.3.2. Gradient boosting model (GBM) 9
2.3.3. Support vector classifier (SVC) 11
2.4. Backward selection of RF and GBM 13
2.5. Confidence interval estimation of an ROC curve 13
2.6. Integrative analysis with data visualization 14
2.6.1. Principal component analysis 14
2.6.2. Principal coordinates analysis 15
2.6.3. Canonical correspondence analysis 16
2.6.4. Co-inertia analysis (CIA) 19
2.7. KEGG analysis 20
3. Result 21
3.1. Subjects and clinical measurement 22
3.2. Sequencing and preprocessing of metagenomic data 22
3.2.1. Shotgun metagenome sequencing 22
3.2.2. Metagenomic data preprocessing and analysis 23
3.3. Annotation of bins 23
3.4. 1H-Nuclear Magnetic Resonance (NMR) spectroscopy and data processing 24
3.5. Diversity of microbiome 24
3.6. Identification of biomarkers for asthma 29
3.7. Integrative analysis 36
3.8. Association between microbiome, metabolite and clinical indices 46
3.9. Function level analysis 48
4. Conclusion and discussion 50
5. Appendix 53
6. Reference 55
7. Weblinks 57

Reference
1. Huang, Y. J. (2015). The respiratory microbiome and innate immunity in asthma. Current opinion in pulmonary medicine, 21(1), 27.
2. Huang, Y. J., Nariya, S., Harris, J. M., Lynch, S. V., Choy, D. F., Arron, J. R., & Boushey, H. (2015). The airway microbiome in patients with severe asthma: associations with disease features and severity. Journal of Allergy and Clinical Immunology, 136(4), 874-884.
3. Hilty, M., Burke, C., Pedro, H., Cardenas, P., Bush, A., Bossley, C., ... & Moffatt, M. F. (2010). Disordered microbial communities in asthmatic airways. PloS one, 5(1).
4. Goleva, E., Jackson, L. P., Harris, J. K., Robertson, C. E., Sutherland, E. R., Hall, C. F., ... & Leung, D. Y. (2013). The effects of airway microbiome on corticosteroid responsiveness in asthma. American journal of respiratory and critical care medicine, 188(10), 1193-1201.
5. Liu, R., Hong, J., Xu, X., Feng, Q., Zhang, D., Gu, Y., ... & Xia, H. (2017). Gut microbiome and serum metabolome alterations in obesity and after weight-loss intervention. Nature medicine, 23(7), 859.
6. Feng, Q., Liang, S., Jia, H., Stadlmayr, A., Tang, L., Lan, Z., ... & Su, L. (2015). Gut microbiome development along the colorectal adenoma–carcinoma sequence. Nature communications, 6, 6528.
7. Chao, A., Chiu, C. H., & Jost, L. (2010). Phylogenetic diversity measures based on Hill numbers. Philosophical Transactions of the Royal Society B: Biological Sciences, 365(1558), 3599-3609.
8. Lozupone, C., Lladser, M. E., Knights, D., Stombaugh, J., & Knight, R. (2011). UniFrac: an effective distance metric for microbial community comparison. The ISME journal, 5(2), 169-172.
9. Martínez-Muñoz, G., & Suárez, A. (2010). Out-of-bag estimation of the optimal sample size in bagging. Pattern Recognition, 43(1), 143-152.
10. Gordon, G., & Tibshirani, R. (2012). Karush-kuhn-tucker conditions. Optimization, 10(725/36), 725.
11. Teer Braak, C. J. F. (1986). Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology, 67, 60-71.
12. Teer Braak, C. J. F. (1985). Correspondence analysis of incidence and abundance data: properties in terms of a unimodal response model. Biometrics, 859-873.
13. Jolliffe, I. T., & Ringrose, T. J. (2004). Canonical correspondence analysis. Encyclopedia of Statistical Sciences, 2.
14. Dray, S., Chessel, D., & Thioulouse, J. (2003). Co‐inertia analysis and the linking of ecological data tables. Ecology, 84(11), 3078-3089.
15. Cave, A.J.; Atkinson, L.L. Asthma in preschool children: a review of the diagnostic challenges. Journal of the American Board of Family Medicine : JABFM 2014, 27, 538-548, doi:10.3122/jabfm.2014.04.130276.
16. Hill, M. O. (1973). Diversity and evenness: a unifying notation and its consequences. Ecology, 54(2), 427-432.
17. Oliveira, G. P., De Abreu, M. G., Pelosi, P., & Rocco, P. R. (2016). Exogenous glutamine in respiratory diseases: myth or reality?. Nutrients, 8(2), 76.
18. Larsen, J. M. (2017). The immune response to Prevotella bacteria in chronic inflammatory disease. Immunology, 151(4), 363-374.
19. Lever, R., & Page, C. (2001). Glycosaminoglycans, airways inflammation and bronchial hyperresponsiveness. Pulmonary pharmacology & therapeutics, 14(3), 249-254.
20. Kim, B. S., Lee, E., Lee, M. J., Kang, M. J., Yoon, J., Cho, H. J., ... & Hong, S. J. (2018). Different functional genes of upper airway microbiome associated with natural course of childhood asthma. Allergy, 73(3), 644-652.
21. Sullivan, A., Hunt, E., MacSharry, J., & Murphy, D. M. (2016). The microbiome and the pathophysiology of asthma. Respiratory research, 17(1), 163.
22. Petchey, O. L., & Gaston, K. J. (2006). Functional diversity: back to basics and looking forward. Ecology letters, 9(6), 741-758.

Weblinks
1. TimeTree: http://www.timetree.org/
2. PHYLIP: http://evolution.genetics.washington.edu/phylip/doc/distance.html