长链非编码RNA识别及其功能挖掘方法研究

Long noncoding RNA (lncRNA) regulates functions of genes and proteins at multiple levels including transcription and translation and is also associated with many diseases. Owing to diversities of structures and types of lncRNA, both the detected disease-related lncRNA and the lncRNA targets account for a little proportion. This seriously blocks understanding of both the cellular mechanisms and the pathological process of diseases. This project is intended to use the sparse representation theory to explore the lncRNA pattern from the multi-source data such as the sequence, structure and expression value, and constructs the deep leaning model to identify lncRNA and its targets. And the project uses lncRNA pattern and the relationships between lncRNA and mRNA, between lncRNA and protein, as well as between lncRNA and miRNA to construct multiple lncRNA-lncRNA functional networks, and exploit the disease-disease similarity network to establish the models to predict lncRNA-disease relationships and to dig lncRNA functions, systematically exploring mechanism of lncRNA-mediated diseases. This study will help find new lncRNA and its targets, help elucidate the mechanism of regulating targets and of lncRNA-mediated diseases, and provide a theoretical foundation for disease prevention and treatment.

长链非编码RNA（lncRNA）在转录和翻译等层面上调控基因和蛋白质的功能，也与很多疾病相关联。由于lncRNA种类与结构的多样性，目前发现的关联疾病的lncRNA及lncRNA靶标仅是很少的一部分，严重制约着对细胞运行机制以及疾病病理过程的认识。本研究拟从序列、结构、表达值等多源数据上采用稀疏表示理论探索lncRNA模式特征，建立识别lncRNA及其靶标的深度学习预测模型；根据lncRNA模式与lncRNA-miRNA、lncRNA-蛋白、lncRNA-mRNA相互作用构建多重lncRNA-lncRNA功能相似性网络，并结合疾病相似性网络构建lncRNA-疾病关联预测及功能挖掘模型，系统地研究疾病在lncRNA调控下的发生机制。本研究将有助于发现新lncRNA及其靶标，有助于阐明lncRNA调控机制和lncRNA所介导的疾病发生机制，为疾病预防与治疗提供理论基础。

人类基因组计划等项目的顺利完成产生了大量的生物数据，为研究人员更深入的探索遗传发育、了解人类疾病致病机理及治疗等提供了海量的数据支持。长链非编码RNA（long noncoding RNAs, lncRNA）是新发现的一族RNA，其变异和失调与人类肿瘤密切相关，探测lncRNA功能将有助于了解lncRNA的调控机制以及深刻地认识疾病或肿瘤发生发展的机制，对人类疾病的预后、诊断和治疗具有十分非常重要的意义。. 本课题结合序列、结构和测序表达等多源数据采用稀疏表示和深度学习理论建立lncRNA及其与靶标的识别模型，探索lncRNA模式及靶标调控机制，构建lncRNA多重功能相似性网络，结合lncRNA与疾病关联网络和疾病功能相似性网络构建全局的lncRNA与疾病的异构网络，探索建立基于复杂网络的lncRNA-disease关联预测模型，从系统上探索疾病在lncRNA、蛋白、miRNA、基因层面的发生机制。.经过一年的研究，本课题组的主要成果有：通过从多源数据上探索lncRNA的模式特征，构建了可信度高的lncRNA与疾病关联权重网络，设计了精确的lncRNA-疾病关联预测及功能挖掘模型，本研究将有助于发现新的1ncRNA及其调控功能，认识lncRNA的产生机制、分子调控机制以及关联lncRNA疾病发生的病理机制，为揭示疾病的发生发展机理及临床诊断与防治提供分子水平上的科学依据，为防治疾病提供理论提供新途径、新方法。