請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/52414
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 郭錦樺 | |
dc.contributor.author | Marisa Huang | en |
dc.contributor.author | 黃文姍 | zh_TW |
dc.date.accessioned | 2021-06-15T16:14:10Z | - |
dc.date.available | 2018-09-24 | |
dc.date.copyright | 2015-09-24 | |
dc.date.issued | 2015 | |
dc.date.submitted | 2015-08-18 | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/52414 | - |
dc.description.abstract | 代謝體學是一個近年來受到許多關注的新研究方法,具有應用於個人化醫療的潛力,但目前此類研究方法仍有一些挑戰。本論文研究臨床代謝體學可能之誤差來源,論文包含兩部分:第一部分針對收集臨床檢體的採血管、採樣體積以及在健康受試者中自體代謝波動等三種變因進行探討。本論文第二部分則為開發一個利用定量代謝體方法針對乳癌進行診斷的分析平台。
在本論文第一部分,我們針對三種常用的血漿(plasma)採樣管,其管內各含不同的抗凝血劑,如肝素(Heparin)、乙二胺四乙酸(EDTA)或是檸檬酸(Citrate),在4mL和8mL的採樣體積下進行分析,並和血清樣品(serum)進行比較。研究發現使用三種採樣管具有不同的結果,並且其中以Heparin管和血清樣品具有較相近的結果;此外,不同的採樣體積對於使用Citrate採樣管以及EDTA採樣管的樣品,有顯著的代謝物濃度差異。對於健康受試者中自體代謝波動分析結果發現,在一個月內四次的不同時間的採樣下,進行健康人的代謝體自體差異分析,結果發現由食物或藥物而導致的代謝物變化有可能高達20倍以上,但內生性的代謝物則在2倍差異以內。 在本論文的第二部分,我們開發了一個利用定量代謝體方法,對於乳癌病患進行血液分析。首先開發了兩個平行的液相層析結合質譜儀 (LC-MS) 的分析平台,並針對52個可能與乳癌相關的代謝物進行分析,比較乳癌患者與健康受試者血液中,此52個代謝物之濃度差異,結果顯示17個代謝物具有偵測乳癌之潛力。因此,本研究後續利用柱後注入內標法 (post-column infused internal standard method) 結合LC-MS方法針對上述17個代謝物進行絕對定量,進而針對停經前後婦女分別建構乳癌預測模型。此模型對於停經前婦女的AUROC為0.940,其專一性及靈敏度分別為94.2%和88.4%;而對於停經後婦女的AUROC為0.828,專一性及靈敏度則分別為85.1%和73.5%。 本論文的結果顯示,使用代謝體研究方法須有合適的實驗設計,才能夠得到正確的結果。本論文同時也建構了一個具有前瞻性診斷乳癌的分析平台。 | zh_TW |
dc.description.abstract | Metabolomics is a field of science that is gaining much popularity in recent years. Metabolomics shows high potential in the area of personalized medicine, yet there remain many challenges within this study approach. This thesis evaluated several challenges in clinical metabolomics. The first part of this study aimed to examine three possible sources of variation seen in metabolomic type studies, including variations that arise from sampling tube types, sampling volumes and also from natural fluctuations within a healthy individual. The second part of this thesis worked on building a quantitative model to diagnose breast cancer using the metabolomic approach.
In the first part, three plasma collection tube types containing the anticoagulants Heparin, EDTA and Citrate were compared with each other and against Serum collection tubes at two different collection volumes (4mL and 8mL). Considerable difference was seen among all tube types, with Heparin tubes displaying the closest results to that of serum tubes. Collection volume greatly influenced the metabolome observed from citrate tubes, and moderately for that of EDTA tubes. Plasma was collected from three healthy individuals four times over the period of a month to see the natural fluctuations of the metabolome under normal conditions. Compounds originating from exogenous sources such as food and drugs fluctuated at extensive folds (>20) whereas the majority of compounds of endogenous origin were regulated within 2 folds. In the second part of the thesis, a quantitative metabolomic method to diagnose breast cancer was developed. Two Liquid Chromatography coupled to Mass Spectrometry (LC-MS) methods were developed to examine 52 selected metabolites in a targeted approach. Breast cancer patients were compared against controls and 17 compounds showed potential to differentiate the two groups. Two post-column infused internal standard LC-MS methods were developed to accurately quantify the select compounds, and applied to a different cohort to build the diagnostic model. The AUROC of the prediction models were 0.940 for premenopausal women, with a specificity of 94.2% and a sensitivity of 88.4%, and 0.828 for postmenopausal women, with a specificity of 85.1% and a sensitivity of 73.5%. Careful consideration for study design must be in place when investigating and interpreting data under metabolomic approach. We present a prospective model for the prediction of breast cancer using a panel of metabolites. | en |
dc.description.provenance | Made available in DSpace on 2021-06-15T16:14:10Z (GMT). No. of bitstreams: 1 ntu-104-R01423026-1.pdf: 2883031 bytes, checksum: ba737224d6c9457b9f305f4e11e4797d (MD5) Previous issue date: 2015 | en |
dc.description.tableofcontents | Table of Contents
PREFACE 10 EVALUATION OF THE SOURCES OF VARIATION IN METABOLOMIC STUDIES INVOLVING LIQUID-CHROMATOGRAPHY COUPLED TO MASS SPECTROMETRY 12 INTRODUCTION 12 MATERIALS AND METHODS 20 Instrumentation 20 Chemicals and Materials 21 Methods 22 INVESTIGATION OF THE EFFECTS OF VARIATIONS IN SAMPLE COLLECTION 25 INVESTIGATION OF THE NATURAL VARIATIONS OF THE METABOLOME 26 SAMPLE TREATMENT 26 RESULTS AND DISCUSSION 27 1. VARIATIONS RESULTING FROM SAMPLING TUBE 27 1.1 INVESTIGATION OF THE EFFECTS OF DIFFERENT ANTI-COAGULANTS ON THE METABOLOME OBTAINED 28 1.2 INVESTIGATION OF THE EFFECTS OF BLOOD TO ANTI-COAGULANT RATIO OF THE METABOLOME OBTAINED 32 1.3 MECHANISM OF ANTICOAGULANT ACTION 37 2. INVESTIGATION OF THE NORMAL DAY TO DAY FLUCTUATIONS IN THE METABOLOME 48 CONCLUSION 61 USING QUANTITATIVE METABOLOMICS TO BUILD A DIAGNOSTIC MODEL FOR THE PREDICTION OF BREAST CANCER 62 INTRODUCTION 62 MATERIALS AND METHODS 70 Instrumentation 70 Chemicals and Materials 71 Methods 72 RESULTS 84 1. USING TARGETED METABOLOMICS TO IDENTIFY DYSREGULATED METABOLITES IN BREAST CANCER PATIENTS 84 1.1 CHOOSING OF THE TARGET METABOLITES 84 1.2LC-MS METHOD DEVELOPMENT 85 1.3 EXTRACTION METHOD OPTIMIZATION 91 1.4 USE TRAINING COHORT TO IDENTIFY DYSREGULATED METABOLITES IN BREAST CANCER PATIENTS 93 2. QUANTIFICATION OF THE DYSREGULATED METABOLITES BY THE PCI-IS METHOD 96 2.1 LC-MS METHOD DEVELOPMENT 96 2.1.1 OPTIMIZATION OF PCI-IS METHOD 98 2.1.2 VALIDATION OF PCI-IS METHOD 102 3. ESTABLISHMENT OF A PREDICTION MODEL TO DETECT BREAST CANCER 106 CONCLUSION 112 REFERENCES 113 LIST OF TABLES Table 1. LC Analysis method for profiling study using Time-Of-Flight (TOF) Mass Spectrometer. 22 Table 2. LC Analysis method for targeted analysis using LC-QqQ. 24 Table 3. Significant Metabolites Displaying Differences Larger than 1.2 Folds Between Different Collection Tubes 30 Table 4. Significant Metabolites Displaying Differences Larger than 1.2 Folds Between Different Collection Volumes 33 Table 5. LC Analysis method for profiling study using Time-Of-Flight (TOF) Mass Spectrometer. 74 Table 6. LC Analysis method for targeted analysis using LC-QqQ. 76 Table 7. LC analysis method for targeted study using LC-QqQ. 78 Table 8 MRM Transitions Used for Targeted Metabolomics Study 79 Table 9. Characteristics of Subjects Included in Training Cohort 93 Table 10 Expected Ranges, PCI-IS Chosen, and MNF Values of Target Analytes 103 Table 11 Equation of the Line and Accuracy of Target Analytes 104 Table 13 Repeatability and Reproducibility of Target Analytes 105 Table 14 Subject Characteristics of Prediction Model Cohort 106 Table 15. Odds Ratios (by multivariate logistic regression) of the respective variables within each model for the four models used to diagnose breast cancer in premenopausal women.. 107 Table 16.Odds Ratios (by multivariate logistic regression) of the respective variables within each model for the four models used to diagnose breast cancer in postmenopausal women.. 108 Table 17 Predictive powers of accepted and under development cancer biomarkers in literature 109 LIST OF FIGURES Figure 1. PCA plot of the metabolome as derived from different sampling tubes at varying sample collection volumes.. 29 Figure 2. Compounds statistically different between different collection volumes in both citrate and EDTA collection tubes 35 Figure 3. Compounds displaying statistically significant differences between varying collection volumes in heparin tubes, with fold changes larger than 1.2 between the tubes. 36 Figure 4. Sites of action of the different anti-coagulants in the sampling tubes discussed in this study. 38 Figure 5. Abundance of nucleoside, nucleotide and analogues analyzed in this study in EDTA tubes at varying collection volumes. 43 Figure 6. Concentration differences between tubes at varying collection volumes of analytes in/or related to the TCA cycle 46 Figure 7. Varying observed concentrations of citric acid from different collection tubes at varying volumes. 47 Figure 8. PCA plot of the metabolomes collected from three healthy volunteers on 4 different days 49 Figure 9. Fold changes between different sampling days of amino acids, peptides and related analogs from the three volunteers 50 Figure 10.Fold changes between different sampling days of lipids and lipid like molecules from the three volunteers 51 Figure 11.Fold changes between different sampling days of carbohydrates, purines and pyrimidines from the three volunteers. 52 Figure 12.Fold changes between different sampling days of organic acids from the three volunteers 53 Figure 13.Fold changes between different sampling days of compounds not belonging to any of the other groupings from the three volunteers. 54 Figure 14. Distribution of compounds analyzed in this study by fold changes. 60 Figure 1. Flow chart of participant selection for the respective sub studies of this research. 83 Figure 2. Example of scoring scheme in relation to peak symmetry 88 Figure 3.Evaluation of the effects of pH on peak shape and abundances of target analytes in the negative mode analytical method as per scoring system described. 89 Figure 4.Mass chromatograms of standard spiked plasma samples analyzed using the a)Negative ionization mode coupled with ion pairing and b) Positive ionization mode coupled with HILIC column 90 Figure 5. Percent abundance of metabolites normalized to abundance achieved using extraction solvent A. 92 Figure 6. Box-plots of the statistically significant compounds in the training cohort. 94 Figure 7.Chromatograms showing improved peak shapes for three amino acids in standard spiked plasma samples. 96 Figure 8. Chromatograms of standard spiked plasma analyzed using the newly optimized a)Negative mode-ion pairing and b)Positive mode ANP methods. 97 Figure 9. Chemical structures of PCI-IS candidates used in this study for both the positive mode and negative mode analytical methods. 100 Figure 10. Chemical structures of the 17 target analytes. 101 Figure 11. ROC curves obtained using each of the four models to diagnose breast cancer in premenopausal women. 107 Figure 12.ROC curves obtained using each of the four models to diagnose breast cancer in postmenopausal women. 108 | |
dc.language.iso | en | |
dc.title | 臨床代謝體樣品收集變異研究與建立偵測乳癌定量模型 | zh_TW |
dc.title | Variations Arising from Sample Collection in Clinical Metabolomics and Quantitative Model Building for the Diagnosis of Breast Cancer | en |
dc.type | Thesis | |
dc.date.schoolyear | 103-2 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 李弘元,陳逸然 | |
dc.subject.keyword | 乳癌,臨床代謝體學,樣品收集,定量代謝體學, | zh_TW |
dc.subject.keyword | Breast Cancer,Clinical Metabolomics,Sample Collection,Quantitative metabolomics, | en |
dc.relation.page | 128 | |
dc.rights.note | 有償授權 | |
dc.date.accepted | 2015-08-18 | |
dc.contributor.author-college | 藥學專業學院 | zh_TW |
dc.contributor.author-dept | 藥學研究所 | zh_TW |
顯示於系所單位: | 藥學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-104-1.pdf 目前未授權公開取用 | 2.82 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。