Computational and Statistical Models for Human Pedigrees

描述（申请人提供）：人类谱系不仅可用于通过连锁分析定位疾病基因，还可用于验证个体间的关系、检测基因分型错误和鉴定单倍型。我们建议建立更快，内存效率更高的谱系分析工具和方法，使遗传学家能够分析数千个单核苷酸多态性（SNP）标记和更大的家庭数据集。我们在这一领域的初步研究已经使遗传学家能够解决比以前可行的更大的问题。例如，我们开发的Merlin程序允许桑格中心和牛津大学的研究人员对7个家系中的1504个22号染色体SNP标记的全染色体单倍型进行表征。该项目的具体目标包括进一步改进家系单倍型的估计，基因型错误的上级检测和建模，以及在考虑多种表型的情况下评估连锁结果的综合模拟框架。我们建议在我们的方法和软件中考虑男女重组率和X染色体数据之间的差异。我们提出的进展将特别有助于那些寻求确定复杂疾病遗传基础的项目，如心血管疾病，糖尿病，肥胖和哮喘。这些疾病通常由多个非独立的表型描述，因此必须通过计算要求高的模拟来评估结果，并且也最有可能依赖于高通量SNP基因分型进行精细定位。我们的工具对于目前正在进行的寻求识别和编目人类基因组中常见单倍型的项目也很重要。我们提出的进展将把许多现在具有挑战性的分析变成常规，并允许研究人员在复杂性状的遗传解剖中提取新的高通量数据源的好处。

DESCRIPTION (provided by applicant): Human pedigrees are useful not only for locating disease genes, through linkage analysis, but also for verifying relationships among individuals, detection of genotyping errors and for identification of haplotypes. We propose to build tools and methods for much faster, memory-efficient pedigree analysis that will enable geneticists to analyze thousands of single-nucleotide polymorphism (SNP) markers and larger family datasets. Our preliminary investigations in this area have already enabled geneticists to tackle larger problems than was previously feasible. For example, Merlin, a program we developed, allowed investigators at the Sanger Center and the University of Oxford to characterize chromosome wide haplotypes for 1504 chromosome 22 SNP markers in 7 pedigrees. Specific aims for this project include further improvements in the estimation of haplotypes in pedigrees, superior detection and modeling of genotype error, and a comprehensive simulation framework for evaluating linkage findings in situations where multiple phenotypes are considered. We propose to allow for differences between male-and-female recombination rates and for X-chromosome data in our methods and software. The advances we propose will be especially useful in projects that seek to identify the genetic basis of complex disease, such as cardiovascular disease, diabetes, obesity and asthma. These diseases are often described by multiple non-independent phenotypes, so that findings must be evaluated through computationally demanding simulation, and are also the most likely to rely on high-throughput SNP genotyping for fine mapping. Our tools will also be important for projects currently underway that seek to identify and catalog common haplotypes in the human genome. The advances we propose will turn many of the analyses that are now challenging into routine and allow investigators to extract the benefits of new high-throughput data-sources in the genetic dissection of complex traits.