基于高维统计因果推断方法检验DNA甲基化对饮酒与上皮性卵巢癌的中介效应

发布时间：2018-03-10 05:10

本文选题：因果推断检验　切入点：表观遗传标记　出处：《山西医科大学》2017年硕士论文　论文类型：学位论文

【摘要】：目的:上皮性卵巢癌(EOC)是美国女性因妇科恶性肿瘤导致死亡的主要原因,同时同时也是恶性肿瘤引起美国女性死亡的第五大原因。EOC的早期诊断缺乏成功,及其所带来的大量机体、社会和经济负担,迫切需要研究者们去研究和阐明EOC的分子基础。关于饮酒与EOC的关系,相关研究结果还未达成一致,部分研究表明饮用果酒可以减小EOC发生的风险。关于饮酒与DNA甲基化的关系,有研究表明饮酒可引起某些位点DNA甲基化改变,而特异位点DNA甲基化改变是EOC发生的早期步骤,可能参与EOC的发病机制。因此如果EOC相关联的DNA甲基化改变是受饮酒调节的话,就为特异DNA甲基化位点及其位置基因作为靶点在EOC预防和治疗中的作用提供依据。本研究旨在发现特异位点DNA甲基化改变作为饮酒与EOC间潜在的中介因素。方法:采用高维因果推断检验方法和VanderWeele中介模型检验饮酒、DNA甲基化和EOC三者的统计因果关联。本研究中采用“去稀疏化”Lasso高维检验方法替代因果推断检验(CIT)过程的第二步检验,将高维检验方法和CIT检验过程相结合应用在表观遗传组学的高维数据分析中。研究对象由Mayo临床卵巢癌研究中196例上皮性卵巢癌病例和202例对照组成。全基因组DNA甲基化数据由398例研究对象外周血白细胞DNA甲基化水平测定。结果:1、CIT检验过程第一步分析结果表明饮酒与EOC有关联,OR=0.336,95%CI[0.169,0.651],P=0.001;第二步分析结果表明控制饮酒变量后,EOC与3,055个CpG位点有关联;第三步分析结果表明控制EOC后,饮酒与61个CpG位点有关联;第四步分析结果表明分别控制2个CpG位点后,饮酒与EOC相互独立。综合四步研究结果,CIT过程共检验出2个CpG位点(cg11016563,cg09358725)作为中介因子,对饮酒-EOC发挥中介效应。2、采用VanderWeele中介模型对CIT检验结果进行交叉验证后,进一步证实2个CpG位点是重要的中介因子,饮酒通过调节2个CpG位点的甲基化水平继而对EOC产生影响。3、cg11016563的基因符号为TRPC6,TRPC6基因过表达与EOC的关联已有研究报道。本文提示TRPC6基因启动子区CpG岛cg11016563位点低甲基化可能通过调节TRPC6基因表达而对EOC产生影响。4、cg09358725的基因符号为LMO2,本文提示LMO2基因可能与EOC有关联。结论:1、首次发现2个CpG位点作为中介因子,饮酒通过调节其甲基化水平继而对EOC产生影响。提示2个CpG位点可能作为EOC发病机制中新的生物标记物。2、TRPC6基因在卵巢癌细胞中的过表达可能受其启动子区CpG岛上cg11016563位点的低甲基化影响。LMO2基因可能与EOC有关联。3、本文提示统计因果推断过程能有效检验环境因素、表观遗传改变和疾病三者间的统计因果关联。4、本文提示高维检验方法与因果推断检验过程相结合可以改进表观遗传组等高维数据的因果推断过程。
[Abstract]:Objective: epithelial ovarian cancer (EOC) is the leading cause of death in women with gynecological malignancies in the United States, and it is also the leading cause of death in women in the United States. The early diagnosis of EOC is not successful. There is an urgent need for researchers to study and elucidate the molecular basis of EOC and the large amount of body, social and economic burdens it brings. There is no agreement on the relationship between alcohol consumption and EOC. Some studies have shown that drinking fruit wine can reduce the risk of EOC. As to the relationship between alcohol consumption and DNA methylation, some studies have shown that drinking alcohol can cause DNA methylation changes at some sites, while DNA methylation at specific sites is the early step of EOC. May be involved in the pathogenesis of EOC. So if the DNA methylation changes associated with EOC are regulated by alcohol consumption, The purpose of this study was to find out that the methylation of specific DNA sites and their location genes may act as potential mediators between alcohol consumption and EOC in the prevention and treatment of EOC. High dimensional causality inference test method and VanderWeele intermediary model were used to test the statistical causal relationship between Lasso methylation and EOC. In this study, the "de-sparse" Lasso high-dimensional test method was used to replace the second step test of causality inference test. The high dimensional test method and the CIT test procedure were used in the analysis of high dimensional data of epigenetics. The subjects were 196 cases of epithelial ovarian cancer and 202 cases of control group in Mayo clinical ovarian cancer study. DNA methylation data were measured by peripheral blood leukocyte DNA methylation level in 398 subjects. Results the first step analysis of EOC test showed that there was a correlation between alcohol consumption and EOC, CI [0.169 卤0.651] P0. 001.The results of the second step analysis showed that after controlling drinking variables, EOC and EOC were higher than that of EOC. 3,055 CpG loci were associated; The results of the third step analysis showed that alcohol consumption was associated with 61 CpG loci after EOC control, and the results of 4th step analysis showed that two CpG loci were controlled respectively. Alcohol consumption and EOC were independent of each other. Two CpG loci (cg11016563cg09358725) were tested by the four-step study results. The alcohol drinking / EOC was used as a mediating effect. The results of CIT test were cross-validated by VanderWeele mediation model. It is further confirmed that two CpG loci are important mediators. Alcohol consumption affects EOC by regulating the methylation level of two CpG loci, and the symbol of the gene. 3cg11016563 is reported to be the association between the overexpression of TRPC6 gene and EOC. This article suggests that the TRPC6 gene promoter region CpG island cg11016563 site is hypomethylated. LMO2 is the gene symbol that may affect EOC by regulating the expression of TRPC6 gene. This paper suggests that the LMO2 gene may be associated with EOC. Conclusion: 1, two CpG loci were first found as mediators. Alcohol consumption affects EOC by regulating its methylation level. It suggests that two CpG loci may act as a new biomarker in the pathogenesis of EOC, and the overexpression of TRPC6 gene in ovarian cancer cells may be affected by the CpG island of its promoter region. The hypomethylation effect of cg11016563 locus. LMO2 gene may be associated with EOC. This paper suggests that the statistical causal inference process can effectively test environmental factors. The statistical causal correlation between epigenetic change and disease. 4. This paper suggests that the combination of high dimensional test method and causality inference test process can improve the causal inference process of high dimensional data such as epigenetic group.
【学位授予单位】：山西医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R737.31

【参考文献】