Population genetics of deleterious polymorphism in human populations

描述（由申请人提供）：尽管有许多单核苷酸多态性（SNP）与疾病的关联研究，但对大多数常见疾病的遗传基础了解甚少。一个可能的原因是，致病的SNP在进化上是不利的，因此，大量的个体罕见突变会导致疾病。使用目前的实验室和统计技术很难研究罕见遗传变异与常见疾病之间的关系。为了设计罕见变异的最佳研究，需要额外的群体遗传模型，其提供罕见和常见SNP之间的相关性或连锁不平衡（LD）的预测。以前开发的有害变异的群体遗传模型不能令人满意地做到这一点。拟议研究的目标是开发和测试预测弱有害SNP之间的LD程度。新模型将结合多个地点的选择和人类人口统计学的现实模型，这两者对于LD模式的预测都至关重要。由于人类人口统计学历史的准确估计对于为新模型提供信息非常重要，因此该项目的第一个目标将开发新方法来从基因组数据中估计人口统计学参数。具体而言，有效使用全基因组重测序数据的马尔可夫链蒙特卡罗方法将允许从复杂的群体遗传模型中精确估计参数。该方法将应用于1000个基因组计划产生的人类群体数据。使用这些参数估计，第二个目标将开发现实的人口遗传模型，可以预测弱有害等位基因周围的LD的程度。这将通过使用简单人口模型的聚结理论进行分析。更复杂的模型将需要模拟。第三个目标将使用来自第三个1000个基因组试点项目的数据，以经验评估不同人群中预测的有害等位基因周围的LD模式。这项工作对于评估和设计罕见变异和疾病风险的研究至关重要。癌症、糖尿病和心脏病等常见疾病是令人难以置信的公共卫生负担。虽然这些疾病都是由遗传风险因素引起的，但在许多情况下，所涉及的特定基因仍然难以捉摸。这项研究将为人类进化史提供新的见解，这对遗传变异和常见疾病风险的设计和解释研究至关重要。

DESCRIPTION (provided by applicant): Despite many association studies of single nucleotide polymorphisms (SNPs) and disease, the genetic basis of most common disease is poorly understood. One possible reason for this is that disease-causing SNPs are evolutionarily disadvantageous, and as a result, a large number of individually rare mutations cause disease. It is difficult to study the relationship between rare genetic variants and common disease using current laboratory and statistical techniques. In order to design optimal studies of rare variants, additional population genetic models, which provide predictions of the correlation, or linkage disequilibrium (LD), between rare and common SNPs are required. Previously developed population genetic models of deleterious variation do not satisfactorily do this. The goal of the proposed research is to develop and test predictions as to the extent of LD between weakly deleterious SNPs. The new models will incorporate selection at multiple sites and realistic models of human demography, both of which are critical for prediction of LD patterns. Since accurate estimates of human demographic history are important for informing the new models, the first Aim of the project will develop novel methods to estimate demographic parameters from genomic data. Specifically, a Markov Chain Monte Carlo approach that efficiently uses genome-wide resequencing data will allow precise parameter estimates from complex population genetic models. This method will be applied to data from human populations generated from the 1000 Genomes Project. Using these parameter estimates, the second Aim will develop realistic population genetic models that can predict the extent of LD around weakly deleterious alleles. This will be done analytically using coalescent theory for simple demographic models. More complex models will require simulation. The third Aim will use data from the third 1000 Genomes Pilot Project to empirically assess patterns of LD around predicted deleterious alleles in different human populations. This work will be critically important for evaluating and designing studies of rare variants and disease risk. Common diseases such as cancer, diabetes and heart disease represent an incredible public health burden. While these diseases are all caused, in part, by genetic risk factors, in many cases, the specific genes involved remain elusive. This research will provide novel insights regarding human evolutionary history which are critical to the design and interpretation studies of genetic variation and risk to common disease.