Bayesian Inference of Haplotypes and Genetic Interactions

描述（由申请人提供）：哺乳动物基因组中最丰富的遗传标记是单核苷酸多态性（SNPs），这些微妙的遗传变异可以是中性的遗传标志，也可以是有害的遗传变化，导致转录失调或蛋白质功能异常。紧密相连的snp等位基因在代代相传时往往会一起移动，这种现象被称为连锁不平衡（linkage disequilibrium， LD）。一条染色体上有序的连锁位点的等位基因配置被称为单倍型。单倍型的独特模式揭示了独特的种群历史，国际单倍图项目描绘的密集基因组尺度SNP图谱有望为研究界提供一个参考标记网格，用于识别单倍型块的共同单倍型，这些单倍型是人类常见疾病遗传易感性的基础。这将大大提高我们开发新的疾病诊断程序和个性化治疗的能力。然而，在过去几年中，随着公共和私人部门SNP基因型数据的积累加速，用于确定单倍型阶段、LD分析和揭示上位相互作用的工具仍然是系统理解和分析人类遗传疾病SNP变异的瓶颈。本研究的总体目标是提高我们分析和理解来自HapMap项目和各种遗传流行病学研究的数据的能力，这些研究利用了HapMap项目产生的密集遗传图谱。更具体地说，我们的目标是(a)开发更强大和准确的算法，从基因型信息推断单倍型阶段，这可以考虑到样本个体之间的进化关系；(b)设计、测试和应用新的统计模型和计算策略，以精细绘制相互作用导致疾病的基因突变；(c)开发、实施和应用新的基于统计模型的算法来检测全基因组关联研究中的基因-基因和基因-环境相互作用。这些任务特别紧迫，因为在很大程度上，我们现在更多地受到我们利用、组织和理解相关遗传数据的能力的限制，而不是产生这些数据的能力。通过在全基因组关联研究中开发有效的单倍型测定、SNP选择、基于ld的精细定位、基因-基因相互作用预测和基因-环境相互作用的算法，我们可以在理解复杂人类性状的遗传基础方面取得巨大进展。

DESCRIPTION (provided by applicant): The most abundantly available genetic markers in the mammalian genomes are single nucleotide polymorphisms (SNPs), and these subtle genetic variations can be either neutral genetic landmarks or detrimental genetic changes resulting in transcriptional dysregulations or protein function abnormalities. Alleles at closely linked SNPs tend to travel together when passed from generation to generation, and this phenomenon has been known as linkage disequilibrium (LD). The allele configuration of an ordered list of linked loci on one chromosome is known as a haplotype. Distinctive patterns of haplotypes reveal distinctive population histories, and the delineation of a dense genome-scale SNP map by the International HapMap project promises to provide the research community with a reference grid of markers for identifying common haplotypes of haplotype blocks that underlie genetic susceptibility to common human diseases. This will greatly enhance our ability in developing novel disease diagnostic procedures and individualized therapies. However, as the accumulation of SNP genotype data accelerates over the past few years in both public and private sections, tools for determining haplotype phases, for LD analysis, and for uncovering epistatic interactions remain a bottleneck towards the systematic understanding and analysis of the SNP variation of the human genetic diseases. The general goal of this research is to advance our capability in analyzing and understanding the data resulting from the HapMap project and various genetic epidemiology studies that take advantage of the dense genetic map produced by the HapMap project. More specifically, we aim to (a) develop more robust and accurate algorithms for inferring haplotype phases from genotype information, which can take into consideration the evolutionary relationship among the sampled individuals; (b) design, test, and apply novel statistical models and computational strategies for fine-mapping genetic mutations that interact to cause diseases; and (c) develop, implement, and apply novel statistical model based algorithms to detect gene-gene and gene-environment interactions in whole-genome association studies. These tasks are particularly urgent because, to a large degree, nowadays we are limited more by our ability in utilizing, organizing, and understanding relevant genetic data than by generating them. By developing effective algorithms for haplotype determination, SNP selection, LD-based fine mapping, gene-gene interaction predictions, and gene-environment interactions in whole genome association studies, tremendous strides can be made in our understanding of the genetic basis of complex human traits.