缺血性脑卒中严重危害健康，风险预测对疾病防治意义重大，包含遗传信息的预测模型在国际上发展迅速，但目前在国内研究较少。本研究拟采用“筛选—验证—预测”的策略，利用通路分析的方法筛选遗传位点，在收集到的家系中进行验证，把遗传、环境危险因素及其交互作用信息结合，在队列人群中建立风险预测模型，探讨中国北方农村人群发病风险评估的科学问题。首先根据国际上已有的全基因组关联研究公共数据库GWASdb的最新信息进行通路分析，挖掘具有生物学意义的位点，在本研究组前期收集的家系中进行连锁和关联分析的重复验证，避免混杂偏倚导致的假阳性；采用多因子降维法等方法探讨交互作用，解决有限样本量分析高维数据交互作用时维度困扰的问题；最后，针对上述发现的阳性位点构建遗传风险评分，利用队列人群建立适合我国农村人群的发病风险预测模型，不仅可以为国内复杂疾病的遗传病因学研究提供有价值的证据，还能实现成果在公共卫生领域的转化应用。

Ischemic stroke can lead to serious health problems, and risk prediction plays a pivotal role in both primary and secondary prevention. Recently risk prediction models with genetic information were widely used, while very little evidence is available in China. Based on the "discovery-validation-application" strategy, this study is designed to screen the genetic Single Nucleotide Polymorphisms (SNPs) by pathway analysis, and then the candidate SNPs will be confirmed in our previous family-based data. Information on genetic, environmental risk factors and their interactions were integrated in the risk prediction models in cohort study in order to improve the ability of risk prediction for rural populations in North China. First, pathway analysis will be implemented in the updated genome-wide association studies (GWAS) public databases such as the GWASdb; candidate genes with biological relevance, will be replicated with linkage and association analyses in our family-based data, to avoid false-positive results due to confounding bias; Methods such as the Multifactor Dimensionality Reduction will be used to characterize the gene-gene and gene-environment interactions, in order to solve the problem on "curse of dimensionality" with limited sample size. Finally, genetic risk scores (GRS) will be constructed with the above positive SNPs, and the GRS predictive properties will be assessed in the prediction model from our established cohort for rural Chinese population. The current study might not only provide more evidence on genetic epidemiology of complex diseases, but also improve public health practice through translational research.