Methods for genomic data with graphical structures

描述（由申请人提供）：该项目的广泛、长期目标涉及开发新的统计方法和计算工具，用于由重要的生物学问题和实验驱动的大规模多基因组学数据的统计和概率建模。新的高通量技术和下一代测序正在产生各种类型的非常高维的基因组和蛋白质组学数据和元数据（例如，网络和途径数据库），以便获得对各种复杂表型的系统级理解。随着数据的数量和复杂性的增加，以及所要解决的问题变得更加复杂，需要能够整合这些基因组数据的统计分析方法，同时可以将有关基因功能和途径的信息纳入数字矢量/矩阵数据的分析中，以便得出有效的统计和生物学推断。当前项目的具体目标是开发新的统计模型和方法，以便在途径和网络的背景下对基因组数据进行综合分析。通过对基因基因组学数据和多种癌症基因组数据的分析，本研究的第一个目标是开发新的统计方法来估计转录水平上一组基因的基因型调整精度矩阵。由此得到的回归系数矩阵和稀疏精度矩阵为调整顺式和转基因对基因表达的影响提供了重要的基因调控信息。第二个目标是开发用于eQTL数据分析的高维工具变量回归，以便在全基因组基因型作为工具变量的情况下确定表型的潜在因果基因。目标3和目标4提出了一套基因集富集分析的新方法，包括通过检验协方差矩阵的同质性来进行基因集分析的方法，以及一类用于综合分析不同基因组数据的多变量随机集方法。这些方法取决于高维回归和高维协方差矩阵估计方法的新颖集成以及先前功能基因集和途径的新颖结合。新方法可应用于不同类型的基因组数据，理想情况下将有助于促进基因及其复杂相互作用的识别，以及各种复杂人类疾病的生物学途径。这里提出的工作将为高维基因组数据建模和研究复杂表型和生物系统提供统计方法，并为各种数据集所代表的每个生物领域提供见解。在此资助下开发的所有项目和详细文件将免费提供给感兴趣的研究人员。

DESCRIPTION (provided by applicant): The broad, long-term objective of this project concerns the development of novel statistical methods and computational tools for statistical and probabilistic modeling of large-scale multiple genomics data motivated by important biological questions and experiments. New high-throughput technologies and next generation sequencing are generating various types of very high-dimensional genomic and proteomic data and metadata (e.g., networks and pathways databases) in order to obtain a systems-level understanding of various complex phenotypes. As the amount and complexity of the data increases and as the questions being addressed become more sophisticated, statistical analysis methods that can integrate these genomic data and in the meanwhile can incorporate information about gene function and pathways into analysis of numerical vector/matrix data are required in order to draw valid statistical and biological inferences. The specific aims of the current project are to develop new statistical models and methods for integrative analysis of genomic data in the context of pathways and networks. Motivated by analysis of genetic genomics data and diverse cancer genomic data, the first aim is to develop novel statistical methods for estimating genotype-adjusted precision matrix for a set of genes at the transcriptional levels. The resulting regression coefficient matrix and sparse precision matrix provide important information on gene regulation when the cis- and trans-genetic effects on gene expressions are adjusted. The second aim is to develop high dimensional instrumental variable regression for eQTL data analysis in order the identify the potential causal genes for a phenotype where the genome-wide genotypes are served as instrumental variables. Aims 3 and 4 propose a set of new methods for gene set enrichment analysis, including methods for gene-set analysis by testing homogeneity of the covariance matrices and a class of multivariate random-set methods for integrative analysis of diverse genomic data. These methods hinge on novel integration of methods for high dimensional regression and high dimensional covariance matrix estimation and novel incorporation of prior functional gene sets and pathways. The new methods can be applied to different types of genomic data and will ideally help facilitate the identification of genes and their complex interactions as well as the biological pathways underlying various complex human diseases. The work proposed here will contribute statistical methodology to modeling high dimensional genomic data and to studying complex phenotypes and biological systems and offer insights into each of the biological areas represented by the various data sets. All programs developed under this grant and detailed documentation will be made available free-of-charge to interested researchers.