STATISTICAL METHODS TO MAP GENES FOR COMPLEX TRAITS

尽管在绘制基因图谱方面取得了巨大的进步 对于孟德尔式疾病来说，影响 对更常见的人类疾病的易感性，特别是 神经精神障碍，进展缓慢的疾病 这些疾病的易感性是非常复杂的，很可能 受到多个基因的影响，这些基因既有微小的影响，也有 相互作用的环境因素。此外，人口 分层，检测疾病的一个主要混杂因素- 标记关联，对解剖提出了特殊的挑战 复杂的特征。需要新的和强大的统计方法来 充分利用海量的宝贵信息 用于过去几年积累的人类关联研究 年，包括信息量很大的标记和密集的地图 那些记号笔。 这项建议的主要目标是扩大和发展 强大的统计方法来检测与潜在基因的联系 对种群分层具有很强抵抗力的复杂特征。 具体地说，我们建议发展：(1)传输不平衡 可以在短时间内联合分析多个标记的测试 在染色体上的物理距离，产生的力量远远大于 单标记方法；(2)统一的框架 适用于更一般情况的传递/不平衡检验 家庭结构比那些可以通过现有的 方法；(3)建模和判别的统计方法 不同的基因-基因互作和基因-环境互作。 理论研究和广泛的计算机模拟都将是 对这些方法进行评估，并与其他方法进行比较 接近了。所开发的方法将应用于可用 用于检测关联和理解复杂病因的数据集 感兴趣的特征。 文档齐全且高效的计算机程序，可实现 新的和强大的方法将被广泛开发和测试 在不同的计算机操作系统下。这些计划将是 通过世界各地向科学界广泛提供 万维网。新的统计方法和强大的计算机 这些计划将为生物医学研究人员提供重要的工具 从海量的基因组中极致最大的信息 为绘制复杂性状的基因图谱而产生的信息 了解这些特征背后的复杂病因。

Despite the tremendous advances that have occured in mapping genes for mendelian diseases, the discovery of genes that influence susceptibility to more common human diseases, particularly the neuropsychiatric disorders, has proceeded slowly The disease susceptibility of these disorders is very complex, and is likely influenced by mulitple genes having mall effects as well as many interacting environmental factors. Furthermore, population stratification, a major confounding factor in detecting disease- marker associations, presents a particular challenge to dissecting complex traits. Novel and powerful statistical methods are needed to take full advangtage of the enormous amount of information invaluable for human linkage studies that has accumulated over the past few years, including both highly informative markers and dense maps of those markers. The major objective of this proposal is to extend and develop powerful statistical methods to detect linkage to genes underlying complex traits that are robust to population stratification. Specifically, we propose to develop: (1) transmission disequilibrium tests that can jointl analyze multiple markers within a short physical distance on a chromosome, yielding far greater power than single-marker methods; (2) a unified framework for the transmission/disequilibrium test that is applicable to more general family structures than those that can be analyzed by existing methods; and (3) statistical methods for modeling and distinguishing different gene-gene interactions and gene-environment interactions. Both theoretical studies and extensive computer simulations will be carried out to evaluate and compare these methods with other approaches. The developed methodology will be applied to available data sets to detect linkage and to understand the complex etiologies of the traits of interest. Well-documented and efficient computer programs that implement the new and powerful methods will be extensively developed and tested under different computer operating systems. These programs will be made widely available to the scientific community through the World Wide Web. The new statistical methods and the powerful computer programs will provide biomedical researchers with important tools to extreact the maiximal information from the enormous amount of genomic information generated to map genes for complex traits and to understand the complex etiologies underlying these traits.