Gene expression and system-based analysis to predict gene-environment interaction

描述（由申请人提供）：尽管许多遗传关联研究的结果已经确定了许多与疾病风险相关的基因变异，但对基因组与环境之间复杂相互作用的认识现在更为普遍。这种基因-环境（GxE）相互作用显示出疾病风险的等位基因特异性改变，并且可能经常通过影响基因表达来响应关键环境因素（EF），如饮食、运动、酒精和烟草的使用。因此，本研究的短期目标是利用生物信息学，根据基因表达数据和基因/蛋白质相互作用网络的分析，以等位基因特异性的方式，优先考虑与饮食成分、身体活动或酒精使用有强烈反应的遗传变异。提出了三个具体目标：一，通过将具有已发表表达QTL的基因与在以特定环境挑战为中心的已发表实验中显示一致表达改变的基因合并，确定假定的GxE相互作用snp，例如高脂肪饮食或热量限制。第二，通过构建基因/蛋白质网络，识别出极有可能表现出GxE相互作用的基因，该网络由携带snp的基因播种，这些snp将等位基因特异性相互作用导向重要的表型或EFs：饮食、运动、饮酒或吸烟。第三，使用目标1和目标2中优先考虑的基因/ snp，测试两个深度表型人群（降脂药物和饮食网络遗传（GOLDN）和弗雷明汉心脏（FHS））中实际的GxE相互作用。虽然GxE相互作用是已知的，并且它们在疾病风险中的作用被认为是更普遍的，但缺乏在广泛的常见环境暴露中快速检测的方法。这里提出的工作意义重大，因为它描述和评估了两种快速有效地优先考虑与心脏病、糖尿病、高血压和肥胖相关的极有可能参与GxE相互作用的遗传变异的方法。目前还缺乏基于计算方法的基因组学数据来预测新的gx。因此，在我们看来，这一提议具有创新性的两个方面是它将系统生物学应用于基因网络和基因表达数据挖掘，以识别对给定EF最敏感的基因，这些基因的变体可能是GxE的参与者。这种创新源于利用基因行为（EF挑战或网络中相互作用伙伴后的表达变化）通过遗传变异（eQTL，基于GxE的网络）过滤，优先考虑GxE相互作用测试的snp。该提案将使用集成基因组学方法来识别假定的GxE变体，这将基于合并大型全基因组数据集并随后过滤以识别具有最多/最佳属性的基因。在这种情况下，eQTL基因为活性基因提供了遗传背景，具有一致mRNA变化的基因是那些对环境线索作出反应的基因，而ef特异性网络中的元素成为进一步分析的候选者，瓶颈是信息流的关键调节器。该研究将在我们的科学家小组中进行，他们精通计算生物学，人类种群遗传学和统计学，并且是世界知名研究机构营养基因组学领域的领导者。此外，我们还获得了两个关键人群的深入表型，用于健康状况的临床测量和饮食、运动和酒精/烟草使用的生活方式选择。

DESCRIPTION (provided by applicant): Although results of many genetic association studies have identified numerous gene variants involved in disease risk, recognition of the complex interaction between genome and environment is now more common. Such gene-environment (GxE) interactions show allele-specific alteration of disease risk and likely often act by affecting gene expression in response to key environmental factors (EF) such as diet, exercise and alcohol and tobacco use. Hence, this study's short- term goal is to use bioinformatics to prioritize genetic variants with a strong likelihood of responding to dietary components, physical activity, or alcohol use in an allele-specific manner based on analysis of gene expression data and gene/protein interaction networks. Three specific aims are proposed: One, identify putative GxE interaction SNPs by merging genes with published expression QTL with genes showing consistent altered expression in published experiments centered on specific environmental challenges, e.g. high-fat diet or caloric restriction. Two, identify genes with strong likelihood to exhibit GxE interactions by building gene/protein networks seeded by genes harboring SNPs directing allele- specific interactions to important phenotypes or EFs: diet, exercise, or alcohol or smoking use. Three, test for actual GxE interactions in two deeply phenotyped populations (Genetics of Lipid Lowering Drugs and Diet Network (GOLDN) and Framingham Heart (FHS)) using genes/SNPs prioritized in Aims 1 & 2. While GxE interactions are known and their role in disease risk is accepted as more commonplace, methods are lacking for rapid detection across a wide range of common environmental exposures. The work proposed here is significant because it describes and assesses two methods for quick and efficient prioritization of genetic variants with high likelihood of partaking in GxE interactions relevant to heart disease, diabetes, hypertension and obesity. Adoption of genomics data to predict novel GxEs based on computational approaches is lacking. Thus, two aspects that, in our opinion, qualify this proposal as innovative are its application of systems biology with gene networks and mining of gene expression data to identify genes most responsive to a given EF, where variants of those genes are likely GxE participants. This innovation arises from leveraging gene behavior (expression changes after EF challenge or interacting partners in a network) filtered through genetic variants (eQTL, GxE- based networks) to prioritize SNPs for the GxE interaction test. This proposal will use integrated genomics methodology to identify putative GxE variants, which will be based on merging large, genome-wide datasets with subsequent filtering to identify the genes with the most/best attributes. In this case, eQTL genes give a genetic context to active genes and genes with consistent mRNA changes are those responding to an environmental cue while elements within the EF-specific networks become candidates for further analysis and bottlenecks are critical regulators of information flow. The proposed research will be performed within our group of scientists who are skilled in computational biology, human population genetics and statistics and are leaders in the field of nutrigenomics at a world-renown research institute. Also, we have access to two key populations deeply phenotyped for both clinical measures of health status and lifestyle choices of diet, exercise and alcohol/tobacco use.