Title Page
Abstract
Contents
Abbreviation 18
Background 21
1. Epigenetics in cancer 21
1-1. What is DNA methylation? 21
1-2. DNA methylation in cancer 22
2. Comparative medicine 26
3. Genome-wide methylome technologies 29
4. Aims of the dissertation 33
CHAPTER Ⅰ. Alternative methylation of intron motifs is associated with cancer-related gene expression in both canine mammary tumor and human breast cancer 35
1. Introduction 36
2. Materials and Methods 39
3. Results 50
4. Discussion 82
CHAPTER Ⅱ. The landscape of PBMC methylome in canine mammary tumors reveals the epigenetic regulation of immune marker genes and its potential application in predicting tumor malignancy 86
1. Introduction 87
2. Materials and Methods 90
3. Results 108
4. Discussion 152
General conclusion 158
Chapter 1 158
Chapter 2 159
References 160
국문초록 169
Background 16
Table B.1. The advantages and disadvantages of representative genome-wide DNA methylation sequencing techniques. 32
Chapter 1 16
Table 1.1. Information for CMT tissue samples used for MBD-seq 40
Table 1.2. Information for CMT tissue samples used for BS-seq 47
Table 1.3. Primers designed for BS-conversion PCR 48
Table 1.4. Quality check for MBD-seq data 51
Chapter 2 17
Table 2.1. The information about dog donors providing blood samples used for MBD-seq 92
Table 2.2. The list of primers designed for targeted BS-sequencing 99
Table 2.3. The list of 127DMRs which have high feature importance in BC classifier 101
Table 2.4. The list of hypermethylated DMRs in immune cell type markers (Panglao DB) 124
Table 2.5. The information of unknown dog PBMC donors (used for validation sets of NT classifier) 145
Background 12
Figure B.1. Schematic representation of DNA methylation and its preferential occurrence at CpG site. 24
Figure B.2. Dynamic changes of methylation across the genome-wide CpG region in cancer. 25
Figure B.3. Leveraging pet dogs as a translational model for human clinical trials in oncology. 28
Figure B.4. Comparison of genome-wide DNA methylation technologies. 31
Figure B.5. Conceptual scheme of the dissertation. 34
Chapter 1 12
Figure 1.1. Schematic presentation of genome wide methylation profiling in CMT. 42
Figure 1.2. Visualizing methylation peaks using processed MBD-seq from 11 pairs of CMT and adjacent normal tissues. 53
Figure 1.3. The CpG coverage of genome wide DNA sequence patterns. 55
Figure 1.4. Analytical strategies. 55
Figure 1.5. Identification of differentially methylated regions (DMRs) among the three CMT subtypes and between CMT and adjacent normal. 56
Figure 1.6. Functional annotation of CMT-DMGs. 59
Figure 1.7. The expression level of the top 4 orthologous genes ranked by OncoScore in canine mammary tumor. 62
Figure 1.8. Functional annotation of Subtype-DMGs. 63
Figure 1.9. Intron DMRs may associate with cancer-related genes. 65
Figure 1.10. Adjust thresholds to select distinguished CMT-DMRs for intronic motif analysis. 69
Figure 1.11. Kaplan-Meier plots showed PAX5 and PAX6 expression reversely effect on the survival rate of breast cancer patients. 70
Figure 1.12. PAX motifs are enriched in hyper- and hypo-methylated intron DMRs. 72
Figure 1.13. PAX motifs are enriched in hyper- and hypo-methylated intron DMRs. Validation of intron hypermethylation in the candidate genes, CDH5... 75
Figure 1.14. Validation of individual CG methylation around PAX5 motif regions in CDH5 and LRIG1 genes. 76
Figure 1.15. Conservation of intron DMRs and associating RNA expression in the candidate genes between HBC and CMT. 79
Chapter 2 14
Figure 2.1. Pair-wise comparison for genome-wide PBMC methylome datasets from benign, carcinoma, and normal dogs. 95
Figure 2.2. Quality check and processing MBD-seq data. 109
Figure 2.3. Venn diagram for hyper- and hypo-methylated DMRs. 112
Figure 2.4. Unsupervised and supervised clustering between comparison groups. 115
Figure 2.5. Gene enrichment analysis for DMGs shows differential immune signatures between tumor and normal PBMCs. 118
Figure 2.6. Enriched terms ranked in the Top 3 by combined score according to comparison groups. 120
Figure 2.7. Immune cell markers involved in normal proliferation and activation of B-cells, T-cells, and NK cells are hypermethylated in tumor PBMCs. 122
Figure 2.8. Targeted CpG methylation and expression analysis in representative hypermethylated genes related to immune cell activation. 138
Figure 2.9. A machine learning-based diagnostic two-step classifier discriminating tumor from normal PBMCs followed by carcinoma from benign PBMCs. 141
Figure 2.10. Evaluating the accuracy and predictive performance of the two- step classifier. 144
Figure 2.11. PCA analysis using DMRs involved in the BC classifiers. 149
Figure 2.12. Permutation accuracy importance of DMRs used for modeling the final BC classifier. 150
Figure 2.13. The predictive performance of transcriptome-based two-step classifier. 151