本课题计划利用计算和实验的方法来研究中国汉族人群中调控遗传变异在阿尔茨海默症（AD）中的作用。具体内容：首先利用人脑全基因组mRNA表达研究数据，结合推算的千人基因组计划的基因型数据，进行数学建模和ENCODE数据分析，寻找能够准确的预测AD相关基因表达量的SNP及其组合，包括前期数据确定的133个侯选基因以及AD相关通路中的基因。然后使用来自美国国立卫生研究院的大规模AD病例-对照关联分析中的数据，并结合自行开发的计算机程序，寻找单独或组合的遗传变异与AD诊断之间可能存在的关联。本课题旨在将预测候选AD致病基因高或低表达的变异组合在一起作为遗传标记用于病例对照研究，使特定等位位点组合与AD致病性之间所存在的关联分析更容易被发现和验证，并能针对个人估算其患AD的遗传风险。

The goal of the present proposal is to use both computational and experimental methods to elucidate the contributions of genetic regulatory variants to Alzheimer's disease. Specifically, we propose to: 1) use data from genome-wide human brain mRNA expression studies (eGWAS) in combination with imputation of 1000 Genomes Project genotypes, mathematical modeling, and data from the ENCODE Project to identify individual variants and combinations of variants that accurately predict the expression of 133 previously identified candidate AD gene and genes that function in AD-related biological pathways; 2) use data from large-scale, AD case-control association studies obtained from the US National Institutes of Health in combination with recently developed computer programs to look for associations between individual variants or combinations of variants with AD diagnosis. We predict that using combinations of variants that predict high- or low expression of candidate AD liability gene as genetic markers in case-control studies will significantly enhance our ability to demonstrate associations of specific genes with AD liability and allow us to estimate the genetic risk of developing AD for individuals.