Semiparametric Methods for Gene-environment Interaction

描述（由申请人提供）：基因组科学的最新进展显着改变了环境健康科学研究的景观。高通量基因组数据的收集对于在环境病例对照和队列研究中调查基因和环境在引起人类疾病中的相互作用变得越来越重要。这种高维基因环境数据的分析提出了大量的统计和计算挑战，特别是在调查基因和环境的相互作用。到目前为止，这一领域的统计发展有限。这种方法上的不足已成为有效研究基因及其与环境相互作用在人类疾病中的作用的瓶颈。本提案的目的是通过开发先进的半参数统计方法来分析基因和环境研究的高通量数据，从而满足这一需求。我们计划（1）在病例对照研究中，开发用于遗传因素，环境暴露和多个外部混杂因素联合效应模型的双稳健估计的半参数局部有效方法，（2）开发用于队列和病例对照研究中的多稳健估计的半参数方法，遗传因素和环境暴露之间相互作用的模型，它们对二元疾病结果产生的影响，（3）发展了半参数遗传效应双稳健推断方法，环境相互作用和混杂调整的考克斯比例风险模型的删失生存数据和（4）开发有效的和开放访问的用户友好的算法和统计软件，实现这些方法的目标，自由传播到基因-环境研究社区。此外，我们将评估我们的方法在三个正在进行的GWAS，我们已经参与以及在模拟研究的性能。 公共卫生相关性：该项目将开发发现新基因和基因-环境相互作用的尖端方法，同时有效地结合先验知识。这些方法对公共卫生领域的影响有望通过开发改进的方法来显著，该方法用于在环境病例对照和队列研究中对基因和环境在引起人类疾病方面的相互作用进行强有力的调查。

DESCRIPTION (provided by applicant): Recent advance of genomic sciences has significantly changed the landscape of environmental health science research. Collection of high throughput genomic data has become increasingly important for investigating the interplay of genes and environment in causing human diseases in environmental case-control and cohort studies. Analysis of such high-dimensional gene-environmental data presents substantial statistical and computational challenges, especially in investigating gene and environment interactions. Limited statistical developments have been made in this area so far. This methodological shortage has become a bottleneck for effectively studying the roles of genes and their interactions with environment in causing human diseases. The purpose of this proposal responds to this need by developing advanced semi-parametric statistical methods to analyze high throughput data from gene and environment studies. We plan (1) to develop semi-parametric locally efficient methods for double-robust estimation in a case-control study, of a model for the joint effect of a genetic factor, an environmental exposure and multiple extraneous confounding factors, (2) to develop semi-parametric methods for multiple robust estimation in cohort and case-control studies, of a model of interaction between a genetic factor and an environmental exposure in the effect that they produce on a binary disease outcome, (3) to develop semi-parametric methods for double robust inferences of genetic effects incorporating gene-environment interaction and confounding adjustment in a Cox proportional hazards model for censored survival data and (4) develop efficient and open access user-friendly algorithms and statistical software that implement these methods with the goal of disseminating them freely to the gene-environment research community. In addition, we will evaluate the performance of our methods in three ongoing GWAS we have been involved with as well as in simulation studies. PUBLIC HEALTH RELEVANCE: The proposed project will develop cutting edge methods for discovery of novel genes and gene-environment interaction while efficiently incorporating prior knowledge. The impact of these methods to the field of public health promises to be significant through the development of improved methodology for robust investigation of the interplay of genes and environment in causing human diseases in environmental case-control and cohort studies.