Variance Component Models for Mapping QTLs

描述（由申请人提供）：人类基因组知识的惊人进步，以及标记技术的平行进步，已经超过了复杂性状统计遗传学方法的互补发展。针对这种情况，目前的建议寻求支持我们在定量性状位点（QTL）定位方面的工作，以适应对现代遗传研究的设计和分析具有实际重要性的数据结构和模型。我们修改后的续期申请是为了支持来自行为遗传学研究所（科罗拉多大学）、精神病学研究所（伦敦大学）、惠康信托人类遗传学中心（牛津大学）和统计遗传学中心（密歇根大学）的合作努力。我们继续这种富有成效的合作的目标是进一步扩展方差成分分析的方法，以适应对现代遗传研究的设计和分析具有实际重要性的更一般的数据结构和模型，并将这些集成到一个全面的软件包中。具体来说，我们将：1)进一步开发对选定样本进行连锁和关联分析的方法，这些方法不仅稳健且无偏倚，而且足以有效地应用于一般谱系；2)利用最优选择程序建立上位性和基因-环境相互作用模型，以检测这些相互作用效应；3)将Fulker et al（1999）将基因型效应划分为兄弟姐妹之间和兄弟姐妹内部成分的模型从单位点等位基因关联扩展到多位点单倍型关联；4)开发能够适应基因分型错误的方差成分方法；5)通过使用回归（Fulker等人，1995）改进多点IBD计算方法，以结合从我们的软件包MERLIN获得的精确的单标记IBD信息；6)将MERLIN扩展为一个综合方差成分分析包，用于对多变量离散和连续数据进行组合、多点联动和关联分析，包括选择样本和随机样本。

DESCRIPTION (provided by applicant):Spectacular advances in knowledge about the human genome, and parallel advances in marker technology, have out paced the complementary development of methods in statistical genetics for complex traits. In response to this situation, the current proposal seeks support of our work on quantitative trait loci (QTL) mapping to accommodate data structures and models that are of practical importance to the design and analysis of modern genetic studies. Our revised renewal application is to support a collaborative effort from the Institute for Behavioral Genetics (University of Colorado), the Institute of Psychiatry (University of London), the Wellcome Trust Centre of Human Genetics (University of Oxford), and the Center for Statistical Genetics (University of Michigan). Our goal in the continuation of this productive collaboration is to further extend the methodology of variance components analysis to accommodate more general data structures and models that are of practical importance to the design and analysis of modern genetic studies, and to integrate these into a comprehensive software package. Specifically we will: 1) further develop methods for linkage and association analysis of selected samples that are not only robust and unbiased but also sufficiently efficient for application to general pedigrees; 2) develop models for epistasis and gene-environment (GE) interaction using optimal selection procedures for the detection of these interactive effects; 3) extend the Fulker et al (1999) model that partitions genotypic effects into between-sibships and within-sibship components from single locus allelic associations to multilocus haplotype associations; 4) develop variance-components methods that are able to accommodate genotyping errors; 5) refine a method for multipoint IBD calculations by use of regression (Fulker et al, 1995) to combine exact single-marker IBD information obtained from our software package, MERLIN; and 6) extend MERLIN into a comprehensive variance components analysis package for combined, multi-point linkage and association analysis of multivariate discrete and continuous data, for both selected and random samples.