The population genetics of disease risk and other quantitative traits

项目摘要 人类全基因组关联研究（GWAS）揭示了大多数遗传变异的特征， 兴趣来自于众多基因的遗传差异。然而，我们使用GWAS来了解特质的能力 生物学和疾病病因学仍然受到限制，因为我们仍然不了解GWAS是如何检测的， 基因对性状变异的贡献与基因对性状生物学的重要性有关。 GWAS揭示的高度多源性还有第二个含义，支持适应性变化的概念。 与人类和其他物种中的性状相关联的基因通常是“多基因的”，即，是由等位基因频率的变化 在许多小的影响位点。人们对这种选择模式的预期行为也知之甚少， 然而，这阻碍了我们在基因组数据中搜索其足迹的能力。此外，尽管预计， 多基因适应应该是普遍存在的，有明显的例子，大的影响适应性差异 （“扫荡”）之间的种群和物种，提出了一个问题的条件下，一种模式， 适应比另一个更受欢迎。在这里，我们计划解决这些差距，我们的理解如下：目标1。 什么时候应该在GWAS中突出“重要基因”？我们将结合联合收割机（i）我们的模型从最后一个赠款期， 它从第一原理描述了一个基因座对焦点性状遗传力的贡献如何取决于它的 影响该性状和其他受稳定选择，与（ii）一个模型，描述如何等位基因的影响， 关于性状由它们的直接（例如，顺式）和间接（例如，反式）对网络中基因活性的影响。 通过这种方式，我们将把一个基因对一个焦点性状（和其他性状）的重要性与它对 焦点性状的遗传力。该模型还将生成关于归因于以下因素的遗传力的预测： 一个基因与其表达水平的遗传变异和其选择约束水平有关;这些预测 将测试>49个数量性状。目标2.表型和等位基因是如何应对不断变化的选择的 复杂特征的压力我们将扩展我们的多基因适应模型，发生在复杂的 数量性状经历适应性最佳值的突然转变，以考虑（i）影响焦点性状的等位基因 通常对其他性状具有有害的多效性影响，即，选择发生在一个多维度的特征中， 空间，以及（ii）复杂性状上的选择压力可能比遗传性状的变化更快 在单次移位后平衡的变化，即，在那个时间段内可能会有重复的变化。目标3.当 我们是否应该期待高度多基因适应性反应而不是涉及少量大影响变化的适应性反应（例如， 扫描）？通过将这些不同的适应模式放在同一个建模框架中， 描述适应性反应的多源性和可预测性如何取决于性状、种群和 选择参数这样，我们就可以提供一个非常需要的理论基础来解释 GWAS的发现，指导寻找多基因适应的足迹，并了解是什么决定了 适应的多样性和可预测性。

PROJECT SUMMARY Genome-wide association studies (GWAS) in humans have revealed that heritable variation in most traits of interest arises from genetic differences in numerous genes. Yet our ability to use GWAS to learn about trait biology and disease etiology remains limited by the fact that we still not understand how what GWAS detect— the contribution of a gene to variation in a trait—relates to the importance of the gene to the biology of the trait. The high polygenicity revealed by GWAS has a second implication, supporting the notion that adaptive changes to traits in humans and in other species will often be “polygenic”, i.e., result from changes in allele frequencies at many small effect loci. How this mode of selection is expected to behave also remains poorly understood, however, impeding our ability to search for its footprints in genomic data. Moreover, despite the expectation that polygenic adaptation should be ubiquitous, there are notable examples of large effect adaptive differences (“sweeps”) between populations and species, raising the question of the conditions under which one mode of adaptation is favored over another. Here, we plan to address these gaps in our understanding as follows: Aim 1. When should ‘important genes’ stand out in GWAS? We will combine (i) our model from the last grant period, which describes from first principles how the contribution of a locus to heritability in a focal trait depends on its effects on that trait and on others subject to stabilizing selection, with (ii) a model describing how allelic effects on traits arise from their direct (e.g., cis) and indirect (e.g., trans) effects on the activity of genes in a network. In this way, we will relate the importance of a gene to a focal trait (and to others) with its contribution to heritability in the focal trait. The modeling will also generate predictions about how the heritability attributed to a gene relates to heritable variance in its expression levels and to its level of selective constraint; these predictions will be tested for >49 quantitative traits. Aim 2. How do phenotypes and alleles respond to changing selection pressures on complex traits? We will extend our modeling of the polygenic adaptation that occurs after a complex quantitative trait experiences a sudden shift in fitness optimum to consider (i) that alleles affecting a focal trait often have deleterious, pleiotropic effects on other traits, i.e., that selection occurs in a multi-dimensional trait space, and (ii) that selection pressures on complex traits may change more rapidly than it takes for genetic variation to equilibrate after a single shift, i.e., that there may be repeated shifts in that timeframe. Aim 3. When should we expect a highly polygenic adaptive response versus one involving few changes of large effect (e.g., sweeps)? By placing these different modes of adaptation within the same modeling framework, we will characterize how the polygenicity and predictability of the adaptive response depend on trait, population and selection parameters. Thus, we will provide a much-needed theoretical foundation with which to interpret GWAS findings, guide the search for the footprint of polygenic adaptation, and understand what determines the polygenicity and predictability of adaptation.