Statistical Methods for Genetic Epidemiology

遗传流行病学统计方法

• 批准号: 8534549
• 负责人: Li Hsu
• 金额: $ 16.82万
• 依托单位: FRED HUTCHINSON CANCER RESEARCH CENTER
• 依托单位国家: 美国
• 资助国家: 美国
• 起止时间: 至
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/8534549
• 关键词: Address; Age; Algorithms; Arts; Beryllium; Case-Control Studies; Clinical; Computing Methodologies; Copy Number Polymorphism; Coronary heart disease; Crohn' s disease; Data; Detection; Development; Diabetes Mellitus; Disease; Disease Association; Disease Outcome; Dose; Environmental Risk Factor; Epidemiologic Studies; Etiology; Event; Face; Family; Genes; Genetic; Genetic Variation; Incidence; Individual; Intervention; Medical; Methods; Modeling; Outcome; Pathway interactions; Prevention strategy; Preventive Intervention; Process; Public Health; Randomized; Recording of previous events; Relative (related person); Research Personnel; Risk; Risk Factors; Sample Size; Statistical Methods; Structure; Testing; Time; Variant; base; cancer type; case control; cost; disorder prevention; disorder risk; flexibility; gene environment interaction; gene interaction; genetic epidemiology; genetic variant; genome wide association study; high throughput technology; human disease; improved; instrument; method development; new technology; novel; statistics; success; tool

This project aims to develop statistical and computational methods for Identifying and haracterizing SNPs (genes) and environmental factors in the involvement of disease occurrence and progression using state-of-art data that are recently generated by high throughput technologies. The first aim proposes to develop and study likelihood based regression strategies to construct flexible combinations of environmental factors that modify genetic effects in association studies that include binary, continuous and time-to-event data. Boosting and regularized regression strategies are used in structured and unstructured interaction models. The second aim proposes the development of improved test statistics for assessing the association of rare variants with disease risk in sequencing studies by adaptively selecting variants variants to use, and by incorporating genome-wide association studies on additional subjects who are not sequenced to increase power. A third aim is concerned with methods development for estimating age-specific absolute risk of genetic variants and environmental factors from case-control studies, which includes a flexible and efficient estimation for time-varying attributable risk function critical in obtaining unbiased estimators of absolute risk and semi- and non-parametric estimation of composite incidence rate from the family history data on cases and controls. A final aim will develop methods to test and estimate the causal effect of an exposure on a clinical outcome, using genetic variants as instruments. The methods exploit the Mendelian randomization of genetic variants and the dose correspondence of genetic effects on the exposure and the the outcome. The project draws on the strength of the studies in which the five investigators are directly involved and addresses the needs and barriers that these studies face. The methods and tools developed in this project will have a broad application to large-scale genetic epidemiologic studies.

该项目旨在开发统计和计算方法，以识别和刺激性SNP （基因）和环境因素参与疾病发生和进展 最近由高通量技术生成的数据。第一个目的提议开发和 基于研究可能性的回归策略，以构建环境因素的灵活组合 在关联研究中修改遗传效应，包括二进制，连续和事件时间数据。 增强和正则回归策略用于结构化和非结构化交互模型。 第二个目的提出了改进的测试统计数据来评估罕见的关联 通过自适应选择用于使用的变体，并通过 对未测序增加的其他受试者进行整合基因组的关联研究 力量。第三个目标是涉及用于估计特定年龄特定绝对风险的方法开发 病例对照研究的遗传变异和环境因素，其中包括灵活有效的 时间变化的归因风险功能在获得绝对风险的无偏估计量中至关重要 以及从家族史数据中对复合发病率的半估计和非参数估计 和控件。最终目标将开发用于测试和估计暴露对A的因果效应的方法 临床结果，使用遗传变异作为仪器。该方法利用了孟德尔的随机化 遗传变异和遗传作用对暴露和结果的剂量对应关系。这 项目借鉴了五名研究人员直接参与的研究的实力，并且 解决这些研究所面临的需求和障碍。该项目开发的方法和工具 将在大规模的遗传流行病学研究中广泛应用。