Scaling up coalescent linkage disequilibrium mapping

描述（由申请人提供）：家族图谱研究通常能够在5厘米的区域内定位疾病基因，但这样的区域可能包含50多个基因。基于种群数据中的连锁不平衡（LD）的方法有可能查明疾病相关基因。本提案从现有的LD制图算法开始，该算法在计算上仅限于0.5 cM或更小的区域，并开发了三种方法，使其可用于人类疾病制图研究。(1)简化重组模型，忽略相邻snp之间的精细重组结构，跟踪较少的重组。(2)沿着染色体在重叠窗口中构建图谱，而不是试图同时分析整个区域。(3)预先计算一个基因组区域的古代谱系关系（其中一个ENCODE区域将用作概念证明）。从本质上讲，所有现代样本都将共享其深层谱系的部分；预计算将大大减少不同群体在同一染色体区域寻找疾病位点所做的冗余工作。这一建议将把一个功能强大但计算代价昂贵的映射算法转化为一个实际使用的算法。发现导致疾病发展的特定基因对诊断、理解和治疗都很重要。建立在疾病基因位置的粗略概念上的诊断测试通常只适用于它们被开发出来的种族；根据实际致病基因或基因进行的检测可以适用于所有人群。对致病基因的了解也可以阐明疾病的机制，并为治疗设计提供靶点。公共卫生相关性：找到导致人类疾病的精确基因或基因对诊断、理解和治疗非常重要。以家庭为基础的研究通常确定一个包含50多个基因的大染色体区域；需要以人口为基础的研究来进一步缩小范围。这一提议将扩展基于人口数据的精细基因定位算法，以便它可以用于更大规模的研究和更广泛的不确定领域。

DESCRIPTION (provided by applicant): Family-mapping studies are often able to locate a disease gene within an area of 5 cM, but such areas may contain 50+ genes. Methods based on linkage disequilibrium (LD) in population data have the potential to pinpoint disease-associated genes. This proposal begins with an existing LD mapping algorithm which is computationally limited to areas of 0.5 cM or less, and develops three approaches to making it usable for human disease-mapping studies. (1) Simplify the model of recombination, tracking fewer recombinations by disregarding fine recombinational structure between adjacent SNPs. (2) Construct the map in overlapping windows along the chromosome, rather than attempting to analyze the entire region simultaneously. (3) Pre-compute the ancient genealogical relationships for a region of the genome (one of the ENCODE regions will be used as a proof of concept). Essentially all modern samples will share portions of their deep genealogy; pre-computation will greatly reduce the redundant work done by different groups seeking disease loci in the same chromosomal region. This proposal will transform a powerful but computationally expensive mapping algorithm into one of practical use. Finding the specific genes which contribute to development of a disease is important in diagnosis, understanding, and treatment. Diagnostic tests built on a rough idea of a disease gene's location often work only in the ethnicity for which they were developed; tests informed by the actual causative gene or genes can work in all populations. Knowledge of the causative genes can also illuminate the mechanisms of disease and provide targets for treatment design. Public Health Relevance: Finding the precise gene or genes contributing to a human disease is important for diagnosis, understanding, and treatment. Family-based studies often identify a large chromosomal region containing 50+ genes; population-based studies are needed to narrow the location further. This proposal will extend a fine-scale gene-location algorithm based on population data so that it can be used in larger studies and across wider areas of uncertainty.