Computational methods and integrative approaches to predict 3D structures of large RNA molecules

核糖核酸（Ribonucleic acids，RNA）是一种具有多种生物学功能的生物分子。例如，作为信使RNA，它编码基因，作为微小RNA，它调节基因，作为核糖体RNA，它翻译基因。为了实现其非编码功能，RNA通常使用可以在两个水平上描述的复杂结构。首先，二级结构包括由典型碱基配对相互作用形成的茎和茎环的最大集合（沃森-克里克和沃布尔）。然后，这些二级结构元件通过大量的货车德瓦尔斯接触和特定的氢键组装在一起形成三级（3D）结构。对完整3D结构的描述是理解RNA分子生物学功能的里程碑。我们建议开发一个强大的计算框架来预测大RNA分子的3D结构。我们的建议是围绕3个不同的模块，这将与加拿大和国际理论家和实验学家合作进行阐述。第一个模块的目的是开发算法来挖掘实验确定的RNA 3D结构存储在数据库中，并确定保守的RNA 3D基序。特别是，我们建议寻找一类新的图案，涉及远程相互作用和非沃森-克里克相互作用。然后，我们将扩展我们的RNA-MoIP框架，使用这些信息来预测大RNA分子的3D结构。在第二个模块中，我们将开发综合方法来确定野生型和突变RNA的SHAPE实验的集合RNA结构。在这里，我们建议扩展我们以前的算法RNAmutants引入的动态编程技术，并添加允许使用RNA基序数据集的新功能。最后，在第三个模块中，我们将应用我们的技术对果蝇和人类基因中的调控元件进行3D结构注释。我们将评估mRNA编码区和非编码区的RNA结构的重要性，并研究对基因调控机制的贡献。该提案中开发的技术旨在开发强大而严格的计算工具，以破译RNA结构在基因调控过程中的作用。这些结果可能有助于我们了解几种疾病的分子基础，并最终为我们提供抑制或恢复基因表达途径的关键信息。本提案中开发的所有方法的源代码将公开发布，并通过网络服务器向社区提供。

Ribonucleic acids (RNAs) are versatile biomolecules that are involved a diverse number of biological functions. For example, as messenger RNA it encodes genes, as microRNA it regulates genes and as ribosomal RNA it translates genes. To achieve their non-coding functions, RNAs often use sophisticated structures that can be described at two levels. First, the secondary structure encompasses the maximal set of stem and stem-loops formed by canonical base-pairing interactions (Watson-Crick and Wobble). Then, these secondary structure elements are assembled together via numerous van der Waals contacts and specific hydrogen bonds into the tertiary (3D) structure. The description of the complete 3D structure is a milestone into the understanding the biological function of a RNA molecule. We propose to develop a robust computational framework to predict the 3D structure of large RNA molecules. Our proposal is articulated around 3 distinct modules, which will be conducted in collaboration with Canadian and international theoreticians and experimentalists.***The first module aims to develop algorithms to mine experimentally determined RNA 3D structures stored in databases and identify conserved RNA 3D motifs. In particular, we propose to search for a new class of motifs involving long-range interaction and non-Watson-Crick interactions. Then, we will expand our RNA-MoIP framework to use this information in order to predict 3D structures of large RNA molecules.***In the second module, we will develop integrative approaches to determine RNA structures from collections of SHAPE experiments on wild type and mutant RNAs. Here, we propose to expand the dynamic programming techniques introduced with our previous algorithm RNAmutants and add new features allowing to use RNA motifs data sets.***Finally, in the third module, we will apply our techniques to perform 3D structural annotation of regulatory elements in drosophila and human genes. We will assess the significance of RNA structures in coding and non-coding regions of mRNAs and investigate contribution to gene regulation mechanisms.***The technology developed in this proposal aims to develop robust and rigorous computational tools to decipher the role of RNA structures in gene regulation processes. These results may help us to understand the molecular basis of several diseases and eventually provide us key information to inhibit or restore gene expression pathways.?The source code of all the methods developed within this proposal will be publicly released and accessible to the community through Web servers.**