基于生物数据融合的恶性肿瘤中miRNA-ceRNA复合协同调控模式的研究

It has been known that miRNAs and competitive endogenous RNAs (ceRNAs) are closely associated with oncogenic initiation and progression, so it is a challenge to explore the composite regulation patterns between miRNA and ceRNA.In the subject, we attempt to propose a systematically analyzing strategy to explore the composite regulation patterns between miRNA and ceRNA for malignant tumor, by integrating multidimensional high-throughput biological data, such as paired miRNA-target gene expression profiles, protein interactome and functional information, etc. Here, long non-coding RNAs (lncRNA) and pseudogenes are considered if they are targeted by miRNAs. Firstly, miRNA-ceRNA functional synergistic network is constructed based on co-regulated functional modules, and the specific topological features of known disease miRNAs/ceRNAs are dissected. Then, we systematically analyze the potential mechanisms underlying miRNA/ceRNA synergetic regulations. In addition, the function of lncRNA and pseudogenes is predicted by considering their ceRNA roles.The ultimate goals of our subject are to detect miRNAs/ceRNAs and miRNA-ceRNA functional synergistic modules which play important roles in oncogenesis, and set up a bioinformatic analysis platform. Then, we apply this platform to identify specific cancer-related miRNAs/ceRNAs and miRNA-ceRNA functional synergistic modules by respectively analyzing cancer datasets, including breast cancer, prostate cancer, glioma and so on. These results promote the research on multiple miRNAs and genes involved in human complex diseases, and provide novel insights in understanding of oncogenic initiation and progression.

恶性肿瘤的发生发展与miRNA和竞争性内源RNA（ceRNA）密切相关，研究miRNA与ceRNAs在癌症中的复合型协同调控模式是具有挑战意义的课题。本课题从系统生物学角度，通过整合miRNA和靶基因双重表达谱、互作组和功能组等高通量生物数据，借助共调控的功能模块构建了miRNA-ceRNA复合型功能协同调控网络、研究了协同调控作用的潜在机制、疾病miRNA/ceRNA的拓扑特征、从ceRNA角度预测长的非编码RNA和假基因功能的新方法和新技术，旨在实现挖掘癌症发生发展过程中关键的miRNA/ceRNA及miRNA-ceRNA协同调控模块，建立癌症关联的miRNA/ceRNA及协同调控模块识别的生物信息学分析平台，应用方法于乳腺癌、前列腺癌、胶质瘤等癌症数据，挖掘出疾病特异的miRNA和ceRNA及协同调控模块，对推进人类复杂疾病多miRNA多基因研究、探讨癌症发生发展的机理具有重大意义。

本项目按原计划完成。在项目中，通过整合miRNA-靶基因互作、蛋白质互作和功能信息等高通量生物数据、我们开发了基于功能模块构建miRNA协同调控网络的方法，探究了疾病miRNA的拓扑特征及协同调控作用的潜在机制，揭示了协同作用的miRNA更倾向受相同转录因子调控，序列更为相似等。另一方面以胶质瘤、乳腺癌、肺癌等20种恶性肿瘤为背景并根据ceRNA互作特点，我们构建了各个癌症相关的ceRNA-ceRNA互作网络，发现大部分ceRNA互作呈现癌症的特异性，也识别了一些癌症中普遍起作用的、差异的和癌症特异的hub ceRNA，功能分析揭示了ceRNA广泛介导癌症hallmark的调控，挖掘了共享的和特异的miRNA-ceRNA协同调控模块，它们参与癌症发生发展过程中共同和特异功能的调控；特别地，我们还剖析了胶质瘤恶性进展过程中miRNA的关键调控作用借助其调控的靶基因，其中hsa-miR-524-5p作为保护因子在胶质瘤恶性进展过程中参与细胞周期调控，并进一步识别了胶质瘤不同阶段的miRNA预后标记物；在ceRNA互作基础上，我们还提出计算方法识别癌症中失调的ceRNA互作，研究发现获得ceRNA互作与丢失ceRNA互作中的ceRNA差异表达模式分布是显著不同的，其中绝大部分获得ceRNA互作对具有相同差异表达方向；进一步地，我们将协同调控网络构建模型移植到lncRNA，识别了恶性肿瘤中lncRNA间的协同调控作用，发现已知的疾病lncRNA倾向有较多的协同调控作用，并倾向以模块为单位行使功能；另外，我们拓展了本研究工作，探讨了恶性肿瘤中miRNA/lncRNA/mRNA的表观修饰（包括组蛋白修饰和DNA甲基化）的异常改变机制，揭示了乳腺癌及不同亚型中miRNA/lncRNA/mRNA广泛受到表观修饰的调控，尤其是基因间区lncRNA，且多种表观修饰组合对miRNA/lncRNA/mRNA表达改变具有更强的影响，从而对乳腺癌发生的关键通路进行调控；为了推广课题相关成果，一些研究结果搭建成在线免费可用的数据资源。本项目共发表SCI论文14篇，刊发在国内外著名的生命科学杂志《Nucleic Acids Research》、《Human Molecular Genetics》、《Oncotarget》等杂志，同时研究成果也进行国内国际交流和实际应用。