Computational Tools for Mining Large Amounts of ChIP and Gene Expression Data

描述（由申请人提供）：ChIP-seq和ChIP-chip（以下简称ChIPx）是绘制全基因组蛋白质-DNA相互作用（PDI）的强大技术。另一方面，微阵列、外显子阵列和RNA-seq被广泛用于测量基因表达。整合ChIPx和基因表达数据提供了一种强大的方法来研究发育和疾病过程中的基因调控。传统上，由单个实验室进行的ChIPx和基因表达实验主要用于研究特定的生物系统。许多实验室的集体努力产生了大量代表不同生物系统的数据。这些数据包含了大量的信息，每个实验室都没有充分利用这些信息。 该提案旨在开发一套协调的计算，统计和软件工具，使科学家能够综合3000多个公开的ChIPx样本和60，000多个人类和小鼠基因表达谱中的信息，以获得新的发现。该项目将把这些异构数据转化为高通量发现生物背景的工具（即，细胞类型、组织和疾病）。首先，将开发一种名为基因集上下文分析（GSCA）的统计方法。GSCA利用大量的公共基因表达数据来推断生物学背景和疾病，其中一个或多个基因集（即，基因组）被协同激活或失活。其次，基于GSCA，我们将开发一种称为转录因子上下文分析（TFCA）的方法。TFCA发现了转录因子（TF）和基因调控途径的新功能背景。该方法首先通过将自己的ChIPx和基因表达数据与公共ChIPx和Gene Ontology数据集成，将TF的靶基因分类为不同的功能类别。然后，它使用GSCA系统地发现与每个类别的功能相关的生物学背景（包括疾病）。总的来说，GSCA和TFCA将为分析ChIPx和基因表达数据建立一个新的范式。传统方法分析与特定系统相关的数据。在新方法中，人们还利用公共ChIPx和基因表达数据中的丰富信息，将一个系统中的发现扩展到其他生物系统。通过允许人们在原始实验范围之外进行新的发现，并将基因调控途径与疾病联系起来，新方法将显着增加新数据和现有数据的价值。 应用GSCA和TFCA，将一起分析人和小鼠中的3000+ ChIPx样本和60，000+基因表达样本，以系统地绘制TF功能和ChIPx定义的疾病调节途径活性。一些新的预测将得到实验验证。除了创造关于各种疾病的新知识外，这项研究还将提供急需的数据集成和数据挖掘工具，帮助科学家将公开的ChIPx和基因表达数据中的丰富信息转化为新发现，并确定有前途的生物医学研究新领域。

DESCRIPTION (provided by applicant): ChIP-seq and ChIP-chip, hereinafter referred to as ChIPx, are powerful technologies to map genome-wide protein-DNA interactions (PDIs). Microarray, exon array and RNA-seq, on the other hand, are widely used to measure gene expression. Integrating ChIPx and gene expression data provides a powerful approach to study gene regulation both during development and in diseases. Traditionally, ChIPx and gene expression experiments conducted by a single laboratory are mainly used to study a specific biological system. The collective efforts of many labs have resulted in a large volume of data representing diverse biological systems. Jointly, these data contain enormous amounts of information that have not been fully utilized by each individual lab. This proposal aims to develop a coordinated set of computational, statistical and software tools to allow scientists to synthesize information in 3000+ publicly available ChIPx samples and 60,000+ gene expression profiles in human and mouse to make new discoveries. The project will turn these heterogeneous data into a tool for high-throughput discovery of biological contexts (i.e., cell types, tissues and diseases) associated with gene regulatory pathway activities. First, a statistical method named Gene Set Context Analysis (GSCA) will be developed. GSCA utilizes large amounts of public gene expression data to infer biological contexts and diseases in which one or more gene sets (i.e., groups of genes) are coordinately activated or inactivated. Second, based on the GSCA, a method called Transcription Factor Context Analysis (TFCA) will be developed. TFCA discovers novel functional contexts of transcription factors (TFs) and gene regulatory pathways. This method first classifies target genes of a TF into different functional categories by integrating one's own ChIPx and gene expression data with public ChIPx and Gene Ontology data. It then uses GSCA to systematically discover biological contexts (including diseases) associated with the function of each category. Collectively, GSCA and TFCA will establish a new paradigm for analyzing ChIPx and gene expression data. The conventional approach analyzes data tied to a particular system. In the new approach, one also leverages the rich information in public ChIPx and gene expression data to extend findings in one system to other biological systems. By allowing one to make novel discoveries beyond the scope of the original experiments and connect gene regulatory pathways to diseases, the new approach will significantly increase the value of both new and existing data. Applying GSCA and TFCA, 3000+ ChIPx samples and 60,000+ gene expression samples in human and mouse will be analyzed together to systematically map TF functions and ChIPx defined regulatory pathway activ- ities to diseases. Some new predictions will be validated experimentally. In addition to creating new knowledge about a variety of diseases, this research will provide urgently needed data integration and data mining tools to help scientists to translate the rich information in the publicly available ChIPx and gene expression data into new discoveries, and identify promising new areas of biomedical research.