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，用于重要生物学问题和实验激发的基因组数据的统计和概率建模。目前项目的具体目标是开发新的统计模型和方法，用于分析具有图形结构的基因组数据，重点是分析遗传途径和网络的方法，包括开发双样本的非参数途径平滑检验和方差分析问题，以确定两个或多个实验条件之间具有扰动活性的途径，为路径平滑的广义线性模型、考克斯比例风险模型和加速失效时间模型开发组Lasso和组阈值梯度下降正则化估计程序，以识别与各种临床表型相关的路径。这些方法的关键在于谱图理论、多元数据分析的非参数方法和统计学习的正则化估计方法的新集成。这些新方法可以应用于不同类型的基因组数据，并将理想地促进识别各种复杂人类疾病和复杂生物过程的基因和生物途径。该项目还将调查这些方法的鲁棒性，功率和效率，并将其与现有方法进行比较。此外，本计画将发展一套实用可行的电脑程式，以执行所提出的方法，并应用于人类心衰、心脏移植排斥反应及神经母细胞瘤的微阵列基因表现研究的真实的资料，以评估这些方法的效能。这里提出的工作将有助于统计方法与图形结构的基因组数据建模，研究复杂的表型和生物系统和方法的高维数据分析，并提供洞察到每个临床领域所代表的各种数据集，以评估这些新方法。在此资助下开发的所有程序和详细文件将通过万维网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 genomic data motivated by important biological questions and experiments. The specific aim of the current project is to develop new statistical models and methods for analysis of genomic data with graphical structures, focusing on methods for analyzing genetic pathways and networks, including the development of nonparametric pathway-smooth tests for two-sample and analysis of variance problems for identifying pathways with perturbed activity between two or multiple experimental conditions, the development of group Lasso and group threshold gradient descent regularized estimation procedures for the pathway-smoothed generalized linear models, Cox proportional hazards models and the accelerated failure time models in order to identify pathways that are related to various clinical phenotypes. These methods hinge on novel integration of spectral graph theory, non-parametric methods for analysis of multivariate data and regularized estimation methods fro statistical learning. The new methods can be applied to different types of genomic data and will ideally facilitate the identification of genes and biological pathways underlying various complex human diseases and complex biological processes. The project will also investigate the robustness, power and efficiencies o these methods and compare them with existing methods. In addition, this project will develop practical a feasible computer programs in order to implement the proposed methods, to evaluate the performance o these methods through application to real data on microarray gene expression studies of human hear failure, cardiac allograft rejection and neuroblastoma. The work proposed here will contribute both statistical methodology to modeling genomic data with graphical structures, to studying complex phenotypes and biological systems and methods for high-dimensional data analysis, and offer insight into each of the clinical areas represented by the various data sets to evaluate these new methods. All programs developed under this grant and detailed documentation will be made available free-of-charge to interested researchers via the World Wide Web.