DEVELOPING STRATEGIES FOR JOINT GENE-ENVIRONMENT ANALYSIS

描述（由申请人提供）：当今遗传分析的一个关键挑战是解释复杂性状中的大部分表型变异，这些变异可归因于遗传因素，但在多次有效的GWAS后仍无法解释（Goldstein 2009;赫什霍恩2009; Kraft and Hunter 2009）。许多人认为，为了了解复杂表型的遗传学，系统地测试联合基因-环境效应（GxE）和基因-基因（GxG）效应的数据集将是必不可少的。然而，高通量，高密度基因组数据（基因型和测序）的新时代已经引起了严重的计算和统计的联合效应的分析挑战。本项目的目标是为联合基因环境效应的统计分析开发有效的策略。具体目标是（1）确定联合GxE效应的多种分析方法的优点和缺点，以及（2）开发协调荟萃分析的新方法。为了实现目标1，我们将系统地检查和比较传统（例如基于回归）和现代（例如分区，机器学习，信息理论）的病例对照和定量表型分析方法。主要产品将是关于每种方法表现良好的条件的具体指导。为了实现目标2，我们将把最近的联合效应方法扩展到一个元分析框架，并对目前使用的方法进行改进。为了这两个目标，我们还将分析与吸烟有关的真实的数据，从而提高我们对吸烟和成瘾风险的遗传和环境因素的理解。这些对真实的数据的分析将由基于生物信息学的变异优先化指导。因此，这项研究将提供指导和工具，推动联合效应分析领域向前发展。因此，它最终将对我们解释目前无法解释的遗传对复杂性状表型方差的贡献的能力产生重大影响。

DESCRIPTION (provided by applicant): A key challenge for genetic analysis today is to account for the bulk of the phenotypic variance in complex traits that is attributable to genetic factors, but remains unexplained after multiple well-powered GWAS (Goldstein 2009; Hirschhorn 2009; Kraft and Hunter 2009). Many believe that systematically testing datasets for joint gene-environment effects (GxE) and gene-gene (GxG) effects will be essential in order to understand the genetics of complex phenotypes. However, the new era of high throughput, high-density genomic data (genotypes and sequencing) has given rise to serious computational and statistical challenges for the analysis of joint effects. The goal of this project is to develop effective strategies for the statistical analysis of joint-gene environment effects. The specific aims are to (1) identify strengths and weaknesses of multiple analytic approaches to joint GxE effects, and (2) develop new methods for coordinated meta-analysis. To accomplish Aim 1, we will systematically examine and compare traditional (e.g. regression based) and modern (e.g. partitioning, machine learning, information theoretic) analysis methods for case-control and quantitative phenotypes. The primary product will be specific guidance as to the conditions under which each method performs well. To accomplish Aim 2, we will extend recent joint-effects methods to a meta-analytic framework and develop improvements to currently used methods. For both aims, we will also analyze real data related to smoking, thereby improving our understanding of genetic and environmental contributors to smoking and addiction risk. These analyses of real data will be guided by bioinformatics-based variant prioritization. This study therefore will provide both guidance and tools needed to move the field of joint effects analysis forward. As a result, it will ultimately have a significant impact on our ability to account for the currently unexplained genetic contribution to phenotypic variance for complex traits.