Computational Analysis of Short Repetitive Motifs in DNA Sequences

描述（由申请人提供）： 基因调控的各个方面的计算分析，特别是转录因子结合，是一个重要的和众所周知的问题。如果得到妥善解决，它将大大提高我们对疾病的理解，并有助于治疗方法的发展。然而，尽管密集的努力和复杂的模型的应用，DNA序列中的结合基序的识别仍然是难以捉摸的。原始序列可能只携带一部分调控信号，而且通常太短，即使是最敏感的算法也无法检测到。为此目的开发的许多软件工具利用启动子区域中基序的聚类和过度表达，有时将这种方法与其他实验或系统发育信息相结合。然而，事实上，许多短序列似乎在DNA的任何片段中过度代表，至少与完全随机模型相比，这阻碍了可靠的发现。 该提案寻求支持，以开发新的软件，并将其应用于识别、可视化和分析重复的短（约5-25个碱基）简并基序，即短（几百个碱基）和长（整个染色体）DNA序列。我们打算将该软件用于人类和其他基因组，以及我们在生物学和化学方面的合作者感兴趣的序列，试图系统地表征短的过度代表序列。我们将确定这些图案对应的实验证实的转录因子结合共识，研究其系统发育的保守性，并调查其可能与重复家庭。将特别注意基因的上游序列，并将开发工具，用于在全基因组范围内搜索相关的基序布局。 我们的软件将基于经典字符串处理算法的改编，通过将种子元素组合成统计上显著的退化图案，以一种新颖的方式解决不精确匹配问题。除了与我们的合作者进行分析外，我们还将把这些程序与我们已经开发和发布的其他工具一起沿着在公共领域，邀请其他研究人员在他们自己的数据上使用它们。

DESCRIPTION (provided by applicant): Computational analysis of various aspects of gene regulation, transcription factor binding in particular, is an important and well known problem. Adequately addressed, it would greatly improve our understanding of diseases and help with the development of treatments. However, despite the intensive efforts and the application of sophisticated models, the identification of the binding motifs in DMA sequences remains elusive. The raw sequence likely carries only a part of the regulatory signal, and it is often too short and subtle to be detected even by the most sensitive algorithms. Many software tools developed for this purpose exploit the clustering and over-representation of motifs in promoter regions, sometimes combining this method with other experimental or phylogenetic information. However, the fact that many short sequences appear to be over-represented in any segment of DNA, at least in comparison with completely random model, impedes the reliable discovery. This proposal seeks support to develop new software and apply it to the identification, visualization and analysis of repeated short (approximately 5-25 bases) degenerate motifs, in short (a few hundred bases) and long (entire chromosomes) DNA sequences. We intend to use this software on the human and other genomes, as well as on sequences of interest to our collaborators in biology and chemistry, in an attempt to systematically characterize short over-represented sequences. We shall determine which of these motifs correspond to the experimentally confirmed transcription factor binding consensuses, study their phylogenetic conservation and investigate their possible association with repeat families. Special attention will be paid to the upstream sequences of genes, and tools will be developed for a genome-wide search for related motif layouts. Our software will be based on an adaptation of classic string processing algorithms to address the inexact matches in a novel way, by combining the seed elements into statistically significant degenerate motifs. In addition to performing analysis with our collaborators, we will place the programs in the public domain, along with the other tools which we have already developed and published, inviting other investigators to use them on their own data.