人体器官的高通量DNA甲基化数据建模与疾病风险预测方法研究

项目名称： 人体器官的高通量DNA甲基化数据建模与疾病风险预测方法研究

项目编号： No.61471078

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 无线电电子学、电信技术

项目作者： 马宝山

作者单位： 大连海事大学

项目金额： 80万元

中文摘要： DNA甲基化与冠心病等复杂疾病密切相关，获得人体器官的DNA甲基化表达数据对预测该器官的患病风险有重要意义。但在多数情况下，很难直接对人体病变器官采样。本项目的研究避免直接在病变器官上取样，而是利用易于取样的外周血作为替代器官组织来探测心脏的患病风险。本项目运用统计遗传学方法，研究人体器官高通量DNA甲基化数据建模和疾病风险的预测。(1)挖掘不同人体组织器官甲基化数据构成的复杂生物网络，研究基因间相互作用以及高通量数据特征的抽取方法；(2)研究结合PCA和多CPG位点等特征的高通量甲基化数据建模与预测方法；(3)结合复杂疾病遗传易感位点定位算法和SVM建立疾病关联模型并预测疾病风险；(4)开发包含上述数学模型和预测算法的生物信息计算平台，通过互联网实现资源共享。本项目的成果不仅可用于心脏疾病的早期探测，还将为采用非侵入式手段的大规模流行病筛查和多基因复杂疾病的临床治疗提供理论依据。

中文关键词： DNA甲基化；疾病预测；数学建模；致病基因挖掘；统计学习

英文摘要： DNA methylation is closely related to the complex diseases such as coronary heart disease, and it is significant to predict disease risk using DNA methylation level (data) of human tissues. Unfortunately, it is difficult to obtain samples from pathological tissues in many cases. In this investigation, we avoid obtaining samples directly from human tissue and apply easy-to-access peripheral blood samples as surrogate tissue to detect heart disease risk. We use statistical genetics methods to study models of high throughput DNA methylation data and disease risk prediction. (1) Analyze the complex biological networks constructed by DNA methylation data of different human tissues, understand interactions and effects among genes and develop the methods to extract the features from high throughput data; (2) Study the approaches to model and predict high throughput methylation data integrating PCA and multiple CPG sites; (3) Build the disease association study models and predict disease risk by SVM and identification algorithm for complex disease genetic susceptibility loci; (4) Develop a computational platform including the above mathematical models, prediction methods and share these resource by internet. The results of our project will not only be applied to predict heart disease risk on the early stage, but also provide theoretical evidences for large scale non-invasive epidemiology screening and clinical therapy of multiple genes complex disease.

英文关键词： DNA methylation;disease prediction;mathematical modeling;pathogenic genes mining;statistical learning