METHODS FOR MULTIPOINT ANALYSIS OF COMPLEX PHENOTYPES

我们以前开发的统计方法，旨在划分 数量性状的遗传方差对位于 特定的染色体区域（Goldgar 1990）。 该方法模拟了 作为同一性模式函数的亲缘关系特征值的协方差 和一组限定给定染色体的标记基因座的重组 地区 这种多点IBD方法（MIM）已在初步的 研究将比现有的方法更强大， 在兄弟姐妹对中的个体基因座上通过血统来确定同一性。 描述的研究 在该建议中，设计进一步表征和扩展该 方法以多种方式。 具体来说，我们将使用蒙特-卡罗 模拟技术，以检查该方法对偏离的鲁棒性 并将其功效和稳健性与 同胞对和传统的依赖模型的连锁分析方法。 其他研究旨在将MIM方法扩展到分析 离散的性状，如复杂的疾病表型与遗传 易感性 类似的鲁棒性和功率模拟研究将在 用于分析离散性状和数量性状。 作为基因组计划的一部分， 正在为每个人类染色体开发。 第二个方面 建议是通过模拟研究确定最有效的 利用这些索引标记进行基因组搜索的策略 对于可能影响给定复杂表型的所有基因座。 MIM方法 将在计算机程序中实现，以便于分析 在三个数据集中收集的几种表型，作为分别 资助的并行研究。 待分析的表型包括： 一组家系中与认知能力相关的数量性状 确定阅读障碍;和定量和离散 表型被假设为常见癌症的前兆。

We previously developed statistical methodology designed to partition the genetic variance of a quantitative trait to the effects of loci located in specific chromosomal regions (Goldgar 1990). This method models the covariance of sibship trait values as a function of the pattern of identity and recombination of a set of marker loci defining a given chromosomal region. This multipoint IBD method (MIM) has been shown in preliminary studies to be considerably more powerful than existing methods based upon identity by descent at individual loci in sib-pairs. The studies described in this proposal are designed to further characterize and extend this method in a number of ways. Specifically, we will use Monte-Carlo simulation techniques to examine the robustness of the method to departures from the underlying assumptions and to compare its power and robustness to both sib-pair and traditional model-dependent methods of linkage analysis. Other studies are designed to extend the MIM method to the analysis of discrete traits such as complex disease phenotypes with a genetic susceptibility. Similar simulation studies of robustness and power will be performed for the analysis of discrete traits as for quantitative traits. As part of the GENOME initiative, a set of highly polymorphic index markers is being developed for every human chromosome. A second aspect of this proposal is to determine through simulation studies the most efficient strategy for utilizing these index markers in conducting genomic searches for all loci which may influence a given complex phenotype. The MIM method will be implemented in a computer program to facilitate the analysis of several phenotypes in three data sets collected as part of separately funded concurrent studies. Phenotypes to be analyzed include: a number of quantitative traits related to cognitive abilities in a set of families ascertained for reading disability; and both quantitative and discrete phenotypes hypothesized to be precursors to common cancer.