全基因组乳腺癌DNA甲基化与基因表达关联模式

发布时间：2018-01-11 13:02

本文关键词：全基因组乳腺癌DNA甲基化与基因表达关联模式　出处：《西安电子科技大学》2015年硕士论文　论文类型：学位论文

更多相关文章： 乳腺癌 甲基化 基因表达 关联模式

【摘要】：乳腺癌是一种严重威胁女性健康的恶性肿瘤,其发病率逐年上升并且具有发病者年轻化的趋势,但是其发病机制却并不清晰,因此对于乳腺癌的研究非常必要。近年来的研究发现乳腺癌的发生是一个多因素综合作用累积的结果,作为一种重要的表观遗传机制,DNA异常甲基化在乳腺癌中非常常见,且往往导致乳腺癌关键基因的表达异常,对乳腺癌的发生发展具有关键性的调节作用。目前在乳腺癌的DNA异常甲基化的研究方面,大部分都是针对单个基因或基因不同区域的DNA甲基化模式的研究,对DNA甲基化与基因表达之间关系的研究并不多,也不够深入和详细,更没有从全基因组的角度来对DNA甲基化与基因表达之间的关系进行全面系统的分析。因此,针对目前研究上存在的不足,本文在全基因组的层面上对乳腺癌DNA甲基化与基因表达之间的关联模式进行了探讨,主要取得的创新性成果如下:1.针对全基因组上DNA甲基化与基因表达数据的高维性和关系的复杂性,本文提出了一种基于差异化分析和聚类的DNA甲基化与基因表达关系分析方法。该方法首先应用SAM差异分析方法筛选出差异表达基因和差异甲基化的CpG位点;然后利用AP聚类算法先对差异甲基化的CpG位点根据相似性聚类形成多个甲基化簇,再针对每个甲基化簇对应的基因表达数据利用AP聚类形成多个基因表达簇,即得到多个甲基化簇和多组基因表达簇;最后对相应簇的甲基化数据和基因表达数据进行组合即得到两者之间的多种关联模式。在上述聚类过程中,该方法通过迭代和设置阈值的方式来避免得到过多的聚类,通过取聚类簇的均值作为簇代表模式的方式来降低计算复杂性。同时该方法可以根据需要调整差异分析方法和聚类方法,具有良好的扩展性。2.本研究对全基因组上乳腺癌DNA甲基化与基因表达的关联模式进行分析发现,得到的关联模式具有显著特点。首先,各类别中的患病样本与正常样本之间存在明显差异性,可基本区分开。其次根据患病样本与正常样本的差异性区别得到的八个类别中,DNA甲基化与基因表达的关联模式具有类间显著不同,类内大致相似的趋势。最后,每个大类中,DNA甲基化与基因表达之间的关联模式存在着微小的差异,类内各关联模式的主要区别在于其甲基化水平和基因表达水平值的分布范围。3.实验分析证明了所得到的关联模式具有良好的生物解释性。首先,各关联模式都存在已知乳腺癌关键致病基因的支持,为新的乳腺癌关键基因的发现及其关联模式的研究提供了依据;其次,对各类别的基因集进行生物通路富集性分析发现,各基因集在多个乳腺癌关键通路中存在显著富集,对乳腺癌的发生发展起着重要调控作用。综上,本研究从全基因组的层面上对乳腺癌DNA甲基化与基因表达之间的关系进行了详细研究和分析,初步建立了乳腺癌DNA甲基化与基因表达之间的关联模式,也为乳腺癌的预后、诊断和治疗提供了方法和依据。
[Abstract]:Breast cancer is a serious threat to the health of women malignant tumor, its incidence has increased year by year and the incidence trend of younger, but its pathogenesis is not clear, so the research of breast cancer is very necessary. The study found that in recent years the occurrence of breast cancer is the accumulation of several factors results. As an important epigenetic mechanism, DNA methylation is common in breast cancer, and often leads to abnormal expression of key genes of breast cancer, regulation is the key to the development of breast cancer. At present, in the study of breast cancer DNA methylation, most of them are on a single gene or gene DNA methylation patterns in different regions, not much research on DNA methylation and gene expression of the relationship between the lack of in-depth and detailed, but not from the genomic perspective A comprehensive and systematic analysis on the relationship between DNA methylation and gene expression. Therefore, to solve the problems existing in present study, the associations between patterns in the whole genome level of breast cancer DNA methylation and the expression of genes is discussed, the main innovation of the paper are as follows: 1. according to the whole genome of DNA methylation and gene expression of high dimensionality and complexity of relational data, this paper proposes a method to analyze the expression of the relationship between DNA methylation and gene differentiation and analysis based on clustering. This method firstly applies SAM difference analysis method screened the differential expression of CpG gene and differential methylation; then using AP clustering algorithm the difference of methylation of the CpG locus based on similarity clustering to form a plurality of methylation clusters, then for each cluster corresponding to the methylation of gene expression data formed by AP clustering multiple gene expression As clusters, multiple methylation clusters and multiple gene expression clusters is obtained; finally the corresponding cluster methylation data and gene expression data were combined to obtain a variety of association patterns between the two. In the process of clustering, the method by setting the threshold and iterative way to avoid excessive clustering. By mean of cluster as the cluster representative mode to reduce the computational complexity. At the same time, the method can be adjusted according to the difference analysis method and clustering method, has good scalability.2. the study of association pattern of expression of breast cancer DNA gene methylation and whole genome analysis found that association patterns obtained with notable features. First, there are obvious differences between the various types of the diseased samples and normal samples, separated from the basic area. Then according to the differences among samples and normal samples of the distinction between Eight categories in association mode with DNA methylation and gene expression were significantly different between classes, class a similar trend. Finally, each of the categories, with tiny differences in patterns of association between DNA methylation and gene expression, the main difference is that the relevance model within a class with good biological explanation patterns associated with the distribution range of.3. value of the level of experimental analysis proved that the expression of the methylation level and gene. Firstly, the related models are known to exist in breast cancer key pathogenic genes support, provide a basis for the study for the detection of breast cancer and its relationship to the new key basic mode; secondly, to each of the categories of gene sets for biological pathway enrichment analysis showed that the gene set in a number of key pathways in breast cancer are significantly enriched for the development of breast cancer plays an important role in regulation. To sum up, this research From the whole genome level of the relationship between the expression of breast cancer DNA gene methylation and makes a detailed study and analysis, established the pattern of association between breast cancer and gene expression of DNA methylation, but also for the prognosis of breast cancer, diagnosis and treatment provide a method and basis.

【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：R737.9

【参考文献】