Statistical analysis of RNA-Seq data

描述（由申请人提供）：RNA- seq是一项最近开发的技术，能够提供细胞样本中RNA群体的全面，核苷酸序列水平调查。该项目的目的是开发严格的统计方法和高效的计算机程序，以便对RNA-Seq实验产生的大量数据进行有效分析。具体来说，我们将以以下目标进行研究。目标1：模拟读取率的非均匀性：众所周知，读取率可以根据相同转录本上的读取位置而有很大差异，这种非均匀性会导致表达量化的偏差。我们将对读取率如何依赖于局部序列上下文进行建模，并设计方法来纠正由非均匀率引起的偏差。目标2：推断异构体特异性表达：即使已知同种异构体，如何将配对端数据纳入统计框架以定量推断异构体表达的问题仍然是一个开放的问题。我们将发展必要的统计理论和方法来解决这一重要问题。目标3：剪接连接的定位、比对和检测：我们将设计计算方法来将reads与参考基因组进行定位和比对，并根据比对结果开发剪接连接的检测方法。目标4：异构体的从头推断：先前目标的结果将被整合和扩展，以开发一个用于推断遗传位点中表达的异构体集的统计框架。基于这个框架，我们将设计算法来发现表达的同种异构体集并量化它们的表达。目标5：开发用于RNA-Seq数据分析的软件：我们将创建一个软件应用程序来支持RNA-Seq数据的分析。从原始序列读取作为输入开始，该软件将允许映射到已知的转录本数据库，发现和显示新的转录本或同种异构体，读取可视化和计算同种异构体特异性表达和相关的统计摘要。通过创建统计和计算工具，以便从RNA-seq数据中提取有用的信息，该项目将加速与人类健康有关的许多领域的研究。

DESCRIPTION (provided by applicant): RNA-Seq is a recently developed technology capable of providing comprehensive, nucleotide sequence-level survey of the RNA population in a sample of cells. The purpose of this project is to develop rigorous statistical methods and efficient computer programs that will allow the effective analysis of the massive amount of data produced by RNA-Seq experiments. Specifically, we will conduct research with the following aims. Aim 1: Modeling non-uniformity of read rates: It is known that read rates can vary substantially depending on the position of the reads on the same transcript and that such non-uniformity can induce biases in expression quantification. We will model how the read rate may depend on local sequence context, and design methods to correct for biases caused by non-uniform rates. Aim 2: Inference of isoform-specific expression: Even when the isoforms are known, the issue of how paired-end data can be incorporated into the statistical framework for quantitative inference of isoform expression is an open problem. We will develop the necessary statistical theory and methods to resolve this important issue. Aim 3: Mapping, alignment and detection of splice junctions: We will design computational methods to map and alignment the reads to the reference genome, and will develop methods for the detection of splice junctions based on the alignment results. Aim 4: De Novo inference of isoforms: The results of the previous aims will be integrated and extended to develop a statistical framework for inferring the set of expressed isoforms in a genetic locus. Based on this framework, we will design algorithms to discover the set of expressed isoforms and to quantify their expressions. Aim 5: Development of software for RNA-Seq data analysis: We will create a software application to support the analysis of RNA-Seq data. Starting from raw sequence reads as input, this software will allow the mapping to known transcript databases, discovery and display of new transcripts or isoforms, visualization of reads and computation of isoform-specific expression and associated statistical summaries. By creating the statistical and computational tools to enable extraction of useful information from RNA-seq data, this project will accelerate many areas of research relevant to human health. PUBLIC HEALTH RELEVANCE: Dr. Wong and his lab members will conduct research on several problems related to the analysis of mRNA data produced by massively parallel sequencing technologies. They will develop statistical models for the inference of isoforms and isoform-specific expression. By creating the tools to enable extraction of useful information from RNA-seq data, this project will accelerate many areas of research relevant to human health.