SOFTWARE FOR INTEGRATED LINKAGE AND ASSOCIATION ANALYSIS

复杂疾病的遗传分析为增加有关导致心脏病、癌症、糖尿病和阿尔茨海默病等常见疾病的生物机制的知识带来了巨大的希望。 然而，复杂疾病的遗传分析是具有挑战性的，因为这些疾病可能是由于多种遗传和环境因素的复杂相互作用。由于这种复杂性，常见疾病的研究需要非常大的样本。 需要新的定量方法，通过扩展现有方法并将其与新方法相结合，最大限度地提高从每项研究中获得的信息。 复杂疾病的研究需要使用多种分析方法，并了解每种方法的优点和缺点，以及不同分析方法可以相互补充的方式，以增强对复杂疾病遗传因素的理解。本提案的目标是开发新的连锁和关联分析方法，并将新方法与现有的核心家庭连锁和关联分析方法联合收割机结合成一个软件包。 具体而言，我们建议：1）扩展现有的受影响的同胞对连锁作图软件（Siblink）以允许考虑表型和环境协变量、额外的非连锁基因和基因分型误差; 2）提供Siblink的增强以允许经验p值和功效的无缝估计、疾病基因位置的区间估计、对模型错误指定的鲁棒性的检查和额外同胞的并入;（3）发展在双亲基因分型信息缺失的情况下进行以家系为基础的关联检验的方法;（4）发展在一次分析中结合有双亲和无双亲的联合收割机家系和使用来自扩展家系的多个家系的方法;（5）进行模拟研究以检查在同时连锁和关联研究的背景下的最佳研究设计;（6）对各种条件下分析方法的选择提供指导。 杜克大学人类遗传学中心和密歇根大学目前正在进行的大量常见疾病的遗传学研究为这些新方法在真实的数据中的应用和评估提供了丰富的资源。

The genetic analysis of complex diseases holds enormous promise for increasing knowledge about the biologic mechanisms leading to common diseases such as heart disease, cancer, diabetes, and Alzheimer disease. However, the genetic analysis of complex disease is challenging because these diseases are likely due to a complex interplay of multiple genetic and environmental factors. As a result of this complexity, studies of common diseases require very large samples. New quantitative methods are needed which maximize the information obtained from each study by extending currently available methods and combining them with novel methods. Studies of complex disease require the use of multiple analytic approaches and an understanding of the advantage and disadvantage of each method as well as the ways in which the different analytic methods can complement each other to enhance understanding of genetic factors for a complex disease. The goal of this proposal is to develop new methods of linkage and association analysis, and to combine the new methods with existing linkage and association methods for nuclear families into single software package. Specifically we propose to: 1) extend existing affected-sib-pair linkage mapping software (Siblink) to allow consideration of phenotypic and environmental covariates, additional unlinked genes, and genotyping error; 2) provide enhancements to Siblink to allow for seamless estimation of empirical p-values and power, interval estimates of disease gene location, examination of robustness to model misspecification and incorporation of additional sibs; 3) develop methods for family-based association tests when parental genotyping information is missing; 4) develop methods to combine families with and without parents in a single analysis and for the use of multiple families from extended pedigrees; 5) perform simulation studies to examine optimal study design in the context of simultaneous linkage and association studies; 6) provide guidance as to the analysis method of choice under various conditions. The study of the genetics of a large number of common diseases currently underway at Duke University Center for Human Genetics and the University of Michigan provides a rich resource for application and assessment of these new methods to real data.