Computational Methods for Deep Sequencing Based RBP Binding Motif Characterizatio

描述（由申请人提供）：RNA结合蛋白由于其结合位点识别的灵活性，可以具有100到1000个目标mrna。这些位点可以同时包含序列和结构元素，并且当比较不同RBP的结合基序时，甚至在同一RBP家族成员之间，也会显示出巨大的差异。直到21世纪初，结合位点的表征主要局限于涉及特定RBP和靶基因/结合基序的个体研究。仅在过去十年中，对体内结合基序的大规模鉴定才成为可能；首先使用rip芯片方法，然后使用CLIP和RIP-seq方法。后两种方法结合了深度测序的力量，使我们不仅能够识别更多的结合位点，而且能够进行更精细的绘图。一个迫切需要关注的领域是计算方法的发展，以对这些技术产生的数据集进行全面分析。在本项目中，我们将利用RIP-seq分析两种结肠癌细胞系中5种rbp的靶点。我们将设计并实现一个框架来表征RNA结合蛋白的结合特异性。我们将从这些数据中评估和改进RBP目标识别的方法。最后，我们将建立一个关于rna结合蛋白及其靶基因、结合位点和结合特异性的信息数据库。

DESCRIPTION (provided by applicant): RNA binding proteins can have 100s to 1000s of target mRNAs thanks to flexibility in recognizing their binding sites. These sites can incorporate both sequence and structural elements and show tremendous variation when binding motifs of different RBPs are compared, even among members of the same RBP family. Until early 2000s, characterization of binding sites was mostly restricted to individual studies involving a particular RBP and a target gene/binding motif. Only in the last decade, en masse identification of in vivo binding motifs became possible; first with the RIP-chip approach and then with CLIP and RIP-seq. These last two methods incorporate the power of deep sequencing, allowing us not only to identify more binding sites but also to conduct a more refined mapping. One area that urges attention is the development of computational methods to perform comprehensive analyses of datasets generated by these techniques. In this project we will profile targets for five RBPs with RIP-seq in two colon cancer cell lines. We will design and implement a framework for characterizing the binding specificity of RNA binding proteins. We will evaluate and refine methodology for RBP target identification from such data. Finally, we will develop a database of information about RNA-binding proteins, their target genes, binding sites and binding specificities.