Integrative analysis of genomic and epigenomic architecture in cognitive decline

描述(由申请人提供)：老年人的认知能力下降和阿尔茨海默病(AD)是这一年龄段严重下降的最常见的具体疾病，是巨大的、日益严重的人文和公共卫生问题。这种轨迹是人类的一种特征，带有可遗传成分，但也与非遗传因素有关。我们建议探索全球认知下降表型的结构，以了解已知和新的危险因素的功能特征。由于认知的下降可能不遵循线性趋势，我们将比较多个统计模型的性能，以检验这一特征，从线性混合效应模型开始，转移到越来越复杂的模型，包括惩罚和二次样条线模型。使用在这些努力中确定的最优统计模型，我们将探索这种增强的认知轨迹估计与现有的数据集之间的关系，这些数据包括(1)基因组范围的基因数据和(2)来自相同个体的背外侧前额叶皮质的基因组范围的DNA甲基化图谱。此外，这些对认知衰退的估计还将与同一受试者大脑同一区域的microRNA(MiRNA)图谱的新数据集进行关联检验。此外，将我们的分析限制在与认知功能下降相关的DNA甲基化和miRNA特征上，我们将评估这两种类型的测量中相互作用的证据，这两种测量捕获了大脑转录状态的不同方面。虽然分别检查遗传、甲基组和转录因子很重要，但评估它们如何共同作用影响认知也将产生信息性结果。因此，我们建议使用模型简化技术，如套索，创建一个认知下降的综合模型，该模型结合了非多余的临床、神经病理、遗传、DNA甲基化和miRNA特征，这些特征是认知下降的危险因素。我们将从已知的临床和遗传成分开始，然后考虑在本提案的早期分析中发现的基因组、DNA甲基化和miRNA特征。在整个提案中，我们将产生已知和新的风险因素的集成模型，这些因素与认知能力下降的遗传结构有关，并继而与AD有关。最终，这些模型将为开发新的研究和分析计划提供理想的基础，这些研究和分析计划将推动我自己作为独立调查员的职业生涯。此外，这些研究肯定会产生许多线索，这些研究的追求将通过创造更多的资金机会来促进我的职业生涯。

DESCRIPTION (provided by applicant): Cognitive decline among older people and Alzheimer's disease (AD), the specific condition most frequently responsible for severe decline in this age group, are large and growing humanistic and public health problems. This trajectory is a human trait with a heritable component, but is also associated with non-genetic factors. We propose to explore the architecture of the phenotype of decline in global cognition to inform the functional characterization of both known and novel risk factors. Since the decline in cognition may not follow a linear trend, we will compare the performance of multiple statistical models with which to examine this trait, starting with a linear mixed effects model and moving to models of increasing complexity, including penalized and quadratic spline models. Using the optimal statistical model determined in these efforts, we will explore the relationship between this enhanced estimate of cognitive trajectories with existing sets of data cataloguing (1) genome-wide genotype data and (2) genome-wide DNA methylation profiles of the dorsolateral prefrontal cortex from the same individuals. In addition, these estimates of cognitive decline will also be examined for associations with a novel dataset of microRNA (miRNA) profiles from the same region of the brain in the same subjects. Further, limiting our analyses to DNA methylation and miRNA features associated with cognitive decline, we will assess for evidence of interaction in these two types of measures, which capture different aspects of the transcriptional state of the brain. While examining genetic, methylomic and transcriptomic factors separately is important, informative results will also come from assessing how they all work together to affect cognition. Therefore, we propose to use model reduction techniques, such as LASSO, to create a comprehensive model of cognitive decline that incorporates non-redundant clinical, neuropathologic, genetic, DNA methylation and miRNA features that are risk factors for cognitive decline. We will start with the known clinical and genetic components and then consider genomic, DNA methylation and miRNA traits discovered in earlier analyses in this proposal. Throughout this proposal, we will produce integrated models of known and new risk factors that relate to the genetic architecture of cognitive decline, and secondarily to AD. Ultimately, these models will provide an ideal basis for developing new studies and analytic plans that will drive my own career as an independent investigator. Moreover, many leads will certainly emerge from these studies, the pursuit of which will serve to further my career by creating additional funding opportunities.