Survival genetics methods for detecting sex-dependent genetic effects on Alzheimer’s disease

项目摘要 阿尔茨海默病(AD)是一种进行性神经退行性疾病，受两种基因的影响 以及环境因素。尽管已经有超过50个具有全基因组意义的风险基因座 到目前为止，AD遗传性的很大一部分仍未得到解释。有了最高的- 吞吐量技术，大量的基因数据已经可以用于AD基因 研究。虽然研究利用这些丰富的数据资源并考虑到性别依赖 遗传效应、多个标记的联合效应和AD风险信息(例如，到AD的时间 表型)为发现新的AD基因带来了巨大的希望，为此类研究提供了严格的分析工具 目前仍缺乏分析。大多数统计工具都不能解释遗传异质性。 此外，现有的多标记生存检验大多基于协变量的COX模型 调整。错误指定协变量调整模型可能会导致虚假关联 调查结果。此外，在队列研究中，AD的时间通常是间隔审查的，并受 相互竞争的死亡风险，但目前还没有多标记生存测试可用于处理间隔时间 审查相互竞争的风险数据。以解决现有方法的局限性，并促进 考虑性别相关的遗传因素对AD预后的遗传关联分析 异质性，我们将开发三个基于加性风险的多标记生存检验 模型、加速失效时间模型和区间截尾生存特征模型。我们 将进一步扩展这三个测试，以进行基因-基因/基因-环境相互作用分析。所有的 新的测试可以处理左截断和竞争风险，这是到AD时间中的两个常见问题 分析。新方法将被编入R包，通过 全面的R档案馆网络。此外，我们将把这些方法应用于英国生物库 和ROSMAP数据来搜索与AD相关的基因并测试基因与性别的相互作用。这个 该项目的成功完成将解决当前AD面临的分析挑战 基因研究，并推动遗传关联统计方法的发展 对总体生存结果进行分析。新方法在英国生物库中的应用 ROSMAP数据将为AD的遗传结构提供新的见解，特别是 性别特异性遗传病因学。

Project Summary Alzheimer's disease (AD) is a progressive neurodegenerative disease influenced by both genetic and environmental factors. Although over 50 risk loci with genome-wide significance have been identified to date, a substantial proportion of AD heritability remains unexplained. With the high- throughput technologies, a large amount of genetic data has become available for AD genetic research. While studies utilizing these enriched data resources and considering sex-dependent genetic effects, joint effects of multiple markers, and AD risk information (e.g., time-to-AD phenotype) hold great promise for novel AD gene discovery, rigorous analytical tools for such analysis are still lacking. Most of the statistical tools can't account for genetic heterogeneity. Besides, existing multi-marker survival tests are largely based on the Cox model for covariate adjustment. Mis-specifying the covariate-adjustment model could lead to spurious association findings. Furthermore, time to AD is usually interval censored in cohort studies and subject to the competing risk of death, but no multi-marker survival test is currently available to handle interval censored competing risks data. To address the limitations of existing methods and facilitate genetic association analysis of time-to-AD outcomes considering sex-related genetic heterogeneity, we will develop three multi-marker survival tests based on the additive hazards model, the accelerated failure time model, and interval censored survival traits, respectively. We will further extend these three tests for gene-gene/gene-environment interaction analyses. All the new tests can deal with left truncation and competing risks, two common issues in time-to-AD analyses. The new methods will be programmed into R packages to be disseminated through the Comprehensive R Archive Network. Additionally, we will apply the methods to the UK Biobank and ROSMAP data to search for AD-associated genes and test for gene-sex interactions. The successful completion of this project will address analytic challenges faced by the ongoing AD genetic research, and advance the statistical methodology development for genetic association analysis of survival outcomes in general. The application of the new methods to the UK Biobank and ROSMAP data will provide new insights into the genetic architecture of AD, especially the sex-specific genetic etiology.