Investigating MicroRNAs and Their Regulatory Networks in Glioblastoma

描述(申请人提供)：多形性胶质母细胞瘤(GBM)是人类最常见和最致命的脑肿瘤。它对放射和化疗具有高度的抵抗力。了解其分子机制对于开发更有效的治疗方法至关重要。最近的研究表明，microRNAs(MiRNAs)在GBM的发病机制中起着关键作用。到目前为止，从各种miRNA表达研究中已经在GBM中发现了100多个调控异常的miRNAs，这为我们提供了对GBM中miRNA表达模式的惊人见解。然而，到目前为止，报道的结果并不一致，这给破译GBM中潜在的miRNA调控机制带来了巨大的挑战。因此，对以前的miRNA数据进行系统检查是迫切需要的，而且是可以执行的。在这个项目中，我们将制定创新的策略，在miRNA调控网络的背景下识别与GBM显著相关的功能重要的miRNAs。该项目将首先通过使用混合效应模型整合多项研究的结果来确定miRNAs的优先顺序，然后构建由加权分子组成的GBM特异性调控网络，即GBM miRNAs、GBM基因和人类转录因子(TF)，最后对调控网络进行密集模块搜索(DMS)，以检测GBM调控网络中功能关键的miRNAs。我们提出了三个具体目标。(1)开发一种新的统计综合框架，用于使用混合效应模型对来自多个研究的miRNA表达数据进行荟萃分析。与传统的汇集分析相比，我们将把每个miRNA的所有可能的效应大小整合到一个混合的 效果建模并计算P值作为其对GBM的总体效果大小。(2)开发一种新的计算管道来构建由GBM miRNAs、GBM基因和转录因子组成的GBM特异性miRNA介导的调控网络。(3)开发一种新的密集模块搜索(DMS)算法，用于识别GBM中功能重要的miRNAs。在DMS算法中，模块被定义为一组FFL，每个FFL包括miRNA、基因(S)和转铁蛋白及其调控关系。该项目是一项开创性的工作，旨在建立一个综合和全面的建模框架，以及用于检测复杂疾病中具有重要功能的miRNAs的实用计算方法，并在GBM中进行演示。这个项目的成功完成将极大地提高我们对GBM监管系统的了解，这可能会导致制定有效的预防、诊断和治疗策略。

DESCRIPTION (provided by applicant): Glioblastoma multiforme (GBM) is the most common and lethal brain tumor in humans. It is highly resistant to radiation and chemotherapy. Understanding its molecular mechanisms is critical in order to develop more effective treatments. Recent studies revealed that microRNAs (miRNAs) play critical roles in the pathogenesis of GBM. To date, more than 100 dysregulated miRNAs have been found in GBM from various miRNA expression studies, which provide us with astonishing insights into the patterns of miRNA expression in GBM. However, the results reported so far have been inconsistent, presenting a great challenge in deciphering the underlying miRNA regulatory mechanisms in GBM. Therefore, a systematic examination of previous miRNA data is immediately needed and executable. In this project, we will develop innovative strategies to identify functionally important miRNAs significantly associated with GBM in the context of miRNA regulatory networks. The project will start by prioritizing miRNAs through integrating results from multiple studies using a mixed effects model, then build GBM-specific regulatory networks comprised of weighted molecules, i.e. GBM miRNAs, GBM genes and human transcription factors (TFs), and finally perform dense module search (DMS) of the regulatory networks to detect functionally critical miRNAs in GBM regulatory networks. We propose three specific aims. (1) To develop a novel, statistical integrative framework for the meta-analysis of miRNA expression data from multiple studies using a mixed effects model. Compared to traditional pooled analysis, we will integrate all possible effect sizes of each miRNA into a mixed effects model and calculate a P-value as its overall effect size to GBM. (2) To develop a novel computational pipeline to construct GBM-specific miRNA- mediated regulatory networks consisting of GBM miRNAs, GBM genes, and TFs. (3) To develop a novel dense module search (DMS) algorithm for identifying functionally important miRNAs in GBM. In this DMS algorithm, a module is defined as a set of FFLs, each of which includes miRNA, gene(s), and TF, and their regulatory relationships. This project constitutes a pioneering effort to establish an integrative and comprehensive modeling framework, as well as practical computational methods for detecting functionally important miRNAs in complex diseases and demonstrates it in GBM. Successful completion of this project will greatly enhance our understanding of the regulatory systems in GBM, which will likely lead to the development of effective prevention, diagnosis, and treatment strategies.