Computational Statistic Approaches to Gene-Environment Interaction

描述（由申请人提供）： 抑郁症、2型糖尿病和哮喘等复杂疾病是由遗传和环境风险因素共同引起的。了解这些危险因素将改善许多此类疾病的诊断和治疗。最近的研究结果表明，基因和环境往往以非线性方式相互作用;遗传风险变异可能会增加暴露的脆弱性或不利后果，但对疾病风险本身没有直接影响。在这种模式下，包括环境协变量可以大大增加基因组扫描的能力。然而，由于缺乏评估基因-环境相互作用的方法，遗传关联的研究受到阻碍。由于可能存在大量的假设，我们认为需要对表型进行详细的定义，并对遗传结构进行精确的建模，以设计强有力的研究。我们建议开发统计方法来估计基因-环境相互作用的家庭数据和样本的无关个体在全基因组关联（GWA）研究。为此，我们组建了一个互动和创新的团队，在基因图谱数据分析方法的开发方面拥有良好的记录。我们的方法是基于映射常见的复杂疾病的风险变异，通过结合多个紧密连锁的标志物和环境协变量的信息。此外，我们提出了算法和模拟工具，以估计基因与环境相互作用的强度，并计划复制研究。我们开发的所有工具和方法都将被纳入公开可用的软件中。我们可以从NIMH双极遗传学倡议中获得基因型和表型数据。我们打算使用这个数据集以及模拟数据集和其他GWA数据集来评估和校准我们的方法，估计基因型-表型相互作用和规划复制研究。(End摘要）

DESCRIPTION (provided by applicant): Complex disorders such as depression, type 2 diabetes and asthma are caused jointly by genetic and environmental risk factors. Understanding these risk factors will improve diagnosis and treatment for many such disorders. Recent results suggest that genes and environment often interact in a nonlinear manner; genetic risk variants may increase the vulnerability to or adverse consequences of exposure, but have no direct effect on disease risk on their own. Under this model, including environmental covariates can increase the power of a genome scan considerably. However, studies of genetic association have been hampered by the lack of methods to assess gene-environment interaction. Due to the large number of hypotheses possible, we feel that detailed definitions of phenotypes and precise modeling of genetic architecture are required to design powerful studies. We propose to develop statistical methods to estimate gene-environment interaction both from family data and from samples of unrelated individuals in a genome-wide association (GWA) study. To this end, we have assembled an interactive and innovative team with a proven track record in the development of methods for the analysis of gene-mapping data. Our approach is based on mapping risk variants for common complex disorders by combining information of multiple tightly linked markers and environmental covariates. Furthermore we propose algorithms and simulation tools to estimate the strength of gene-environment interaction and to plan replication studies. All the tools and methods we develop will be incorporated into publicly available software. We have access to genotype and phenotype data from the NIMH bipolar genetics initiative. We intend to use this dataset as well as simulated datasets and other GWA datasets to evaluate and calibrate our methods for estimating genotype-phenotype interaction and for planning replication studies. (End of Abstract)