肿瘤异质性研究的非负矩阵分解模型

发布时间：2018-05-02 23:56

本文选题：DNA甲基化 + 肿瘤细胞纯度　；参考：《上海师范大学》2017年博士论文

【摘要】：癌症是一类由于细胞分裂和凋亡机制失常而导致的疾病,经常表现为恶性肿瘤。肿瘤异质性是恶性肿瘤的特征之一,由于无法全面解析它的亚克隆结构,加上治疗后容易产生抗药性和发生转移,给治疗提出极大的挑战。DNA甲基化作为DNA分子的表观遗传修饰,被认为是肿瘤和其它人类疾病发生的原因。通过DNA甲基化估计肿瘤内部各个亚群的比例有助于了解肿瘤的发展过程,并能提出合理的个体化用药和治疗方案。本文首先介绍了肿瘤纯度估计和差异甲基化分析的若干计算方法。其次,分别利用TCGA数据库中不同类型的分子数据建立了基于非负矩阵分解和二次规划的肿瘤异质性分解的如下几个计算模型:一、基于DNA甲基化的肿瘤异质性分解模型;二、基于基因表达的肿瘤异质性分解模型;三、联合DNA甲基化基因表达的肿瘤异质性分解模型。数据模拟发现,我们的方法可以成功地估计出肿瘤内部各个亚群的比例,并且基于多数据整合的模型相对于单个数据类型具有更高的准确度。作为一个特例,我们的方法对于肿瘤纯度(设定肿瘤亚群落数量为2)的估计也与InfiniumPurify,ABSOLUTE,ESTIMATE等方法具有非常好的一致性。
[Abstract]:Cancer is a kind of disease caused by cell division and apoptosis. Tumor heterogeneity is one of the characteristics of malignant tumor. Because it is unable to fully analyze its subclonal structure, and it is easy to produce drug resistance and metastasis after treatment, it is a great challenge to treat. DNA methylation is an epigenetic modification of DNA molecule. It is thought to be the cause of tumors and other human diseases. Estimating the proportion of each subgroup of tumor by DNA methylation is helpful to understand the process of tumor development and to propose a reasonable individualized drug and treatment scheme. In this paper, we first introduce some calculation methods of tumor purity estimation and differential methylation analysis. Secondly, using different types of molecular data in TCGA database, the following computational models of tumor heterogeneity decomposition based on nonnegative matrix decomposition and quadratic programming are established: first, tumor heterogeneity decomposition model based on DNA methylation; Tumor heterogeneity decomposition model based on gene expression and tumor heterogeneity decomposition model combined with DNA methylation gene expression. The data simulation shows that our method can successfully estimate the proportion of each subgroup in the tumor and the model based on multi-data integration has a higher accuracy than a single data type. As a special case, our method is in good agreement with the methods of Infinium Purifyn ABSOLUTETEESTIMATE and so on.
【学位授予单位】：上海师范大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：R73-3;O151.21

【相似文献】