Methods for high-resolution analysis of genetic effects on gene expression

描述(由申请人提供)：评估遗传变异对细胞表型的影响，如基因表达，为理解基因组和疾病的生物学提供了新的机会。通过鉴定表达和拼接数量性状基因座(eQTL或sQTL)，我们可以阐明性状相关变异的新机制，并获得对基因调控机制和途径的新见解。随着新技术和新的大型数据集的出现，我们现在能够对转录本进行高分辨率分析，并阐明表型变异和疾病的因果细胞机制。在拟议的项目中，我们的目标如下：具体目标1：我们将对GTEx数据进行详细的转录组分析。我们将通过采用新的计算方法来改进转录组分析的工作流程。首先，我们将解决识别抄本和估计其丰度的问题。其次，我们将采用贝叶斯方法来比较人群和组织内部和之间的转录本分布，以建立一个强大的差异表达基因目录。具体目标2：我们将开发改进的统计方法来发现监管变异。在过去的3-4年里，我们的合作小组已经开发了许多工具来绘制个体之间表达差异的遗传变异图谱。在这里，我们将这些工具应用到GTEx数据中，使用Aim 1的高质量转录组特征量化来定位eQTL。我们将采用的方法包括：(I)基于单倍型的eQTL定位方法，(Ii)改进的等位基因特异表达(ASE)量化方法，(Iii)使用GRN对反式eQTL进行贝叶斯定位，以及(Iv)整合多组织和多群体eQTL定位。具体目标3：我们将定位影响基因表达或转录本结构的假定因果变异，并评估它们的功能属性。为了了解人类基因调控的分子基础，我们将创建一个影响表达的因果变异及其相关基因组特征的全面目录。我们将集中于：(I)染色质状态模式的研究，以确定eQTL的位置和效应的规则；以及(Ii)功能变异对转录本和个体的影响的解释。具体目标4：我们将建立细胞转录丰度的定量遗传和基因调控模型。在这一目标下，我们的主要工作将是：(I)评估蛋白质编码和调控变异之间的上位性/外显性模式；以及(Ii)重建基因调控网络。这些模型将为eQTL的原因和后果提供生物学上的见解。

DESCRIPTION (provided by applicant): Assessing the impact of genetic variants on cellular phenotypes like gene expression provide new opportunities for understanding the biology of genomes and disease. By identifying expression and splicing quantitative trait loci (eQTL or sQTL), we can elucidate new mechanisms underlying trait-associated variation and gain new insights into gene regulatory mechanisms and pathways. With the availability of novel technologies and new large datasets we are now in a position to perform high resolution analysis of transcriptomes and elucidate causal cellular mechanisms for phenotypic variability and disease. In the proposed project we aim to do the following: Specific Aim 1: We will undertake detailed transcriptome analysis of the GTEx data. We will improve the workflow of transcriptome analysis by deploying novel computational methods. First, we will tackle the problem of identifying transcripts and estimating their abundances. Second, we will deploy a Bayesian approach for comparing transcript distribution within and among populations and tissues to develop a robust catalog of differentially expressed genes. Specific Aim 2: We will develop improved statistical methods to discover regulatory variation. Over the past 3-4 years, our collaborative group has developed many tools for mapping genetic variants underlying expression differences among individuals. Here, we will apply these tools to the GTEx data to map eQTL using the high-quality transcriptome feature quantifications from Aim 1. The approaches we will deploy include: (i) haplotype-based methods for mapping of cis eQTL, (ii) improved methods for quantifying Allele Specific Expression (ASE), (iii) Bayesian mapping of trans eQTL using GRNs, and (iv) integrated multi-tissue and multi-population eQTL mapping. Specific Aim 3: We will map putatively causal variants that affect gene expression or transcript structure and assess their functional attributes. To understand the molecular bases of human gene regulation, we will create a comprehensive catalog of causal variants influencing expression and their associated genomic features. We will focus on: (i) the study of patterns of chromatin states to define rules for the location and effect of eQTL; and (ii) the interpretation o loss of function variant effects on transcriptomes and individuals. Specific Aim 4: We will build quantitative genetic and gene regulatory models of cellular transcript abundance. Our main efforts under this aim will be: (i) to assess patterns of epistasis/penetrance between protein-coding and regulatory variation; and (ii) reconstruct gene regulatory networks. These models will provide biological insights into the causes and consequences of eQTL.