Rosetta Inpharmatics LLC, A wholly owned subsidiary of Merck & Co, Seattle, Washington, United States of America.
PLoS One. 2010 Dec 14;5(12):e14319. doi: 10.1371/journal.pone.0014319.
To identify the genes and pathways that underlie cardiovascular and metabolic phenotypes we performed an integrated analysis of a mouse C57BL/6JxA/J F2 (B6AF2) cross by relating genome-wide gene expression data from adipose, kidney, and liver tissues to physiological endpoints measured in the population. We have identified a large number of trait QTLs including loci driving variation in cardiac function on chromosomes 2 and 6 and a hotspot for adiposity, energy metabolism, and glucose traits on chromosome 8. Integration of adipose gene expression data identified a core set of genes that drive the chromosome 8 adiposity QTL. This chromosome 8 trans eQTL signature contains genes associated with mitochondrial function and oxidative phosphorylation and maps to a subnetwork with conserved function in humans that was previously implicated in human obesity. In addition, human eSNPs corresponding to orthologous genes from the signature show enrichment for association to type II diabetes in the DIAGRAM cohort, supporting the idea that the chromosome 8 locus perturbs a molecular network that in humans senses variations in DNA and in turn affects metabolic disease risk. We functionally validate predictions from this approach by demonstrating metabolic phenotypes in knockout mice for three genes from the trans eQTL signature, Akr1b8, Emr1, and Rgs2. In addition we show that the transcriptional signatures for knockout of two of these genes, Akr1b8 and Rgs2, map to the F2 network modules associated with the chromosome 8 trans eQTL signature and that these modules are in turn very significantly correlated with adiposity in the F2 population. Overall this study demonstrates how integrating gene expression data with QTL analysis in a network-based framework can aid in the elucidation of the molecular drivers of disease that can be translated from mice to humans.
为了确定心血管和代谢表型的基因和途径，我们通过将来自脂肪组织、肾脏和肝脏组织的全基因组基因表达数据与群体中测量的生理终点相关联，对 C57BL/6JxA/J F2（B6AF2）杂交小鼠进行了综合分析。我们已经确定了大量性状 QTL，包括在染色体 2 和 6 上驱动心脏功能变异的基因座，以及在染色体 8 上驱动肥胖、能量代谢和葡萄糖性状的热点。脂肪基因表达数据的整合确定了一组核心基因，这些基因驱动染色体 8 肥胖 QTL。这个染色体 8 跨表达数量性状基因座（eQTL）特征包含与线粒体功能和氧化磷酸化相关的基因，并且映射到人类中具有保守功能的子网络，该子网络先前与人类肥胖有关。此外，与签名特征中的同源基因相对应的人类 eSNP 在 DIAGRAM 队列中与 II 型糖尿病的关联显示出富集，这支持了染色体 8 位置扰乱了一个分子网络的观点，该网络在人类中感知 DNA 的变化，进而影响代谢疾病风险。我们通过证明三个来自跨 eQTL 特征的基因（Akr1b8、Emr1 和 Rgs2）敲除小鼠的代谢表型，从功能上验证了这种方法的预测。此外，我们还表明，这两个基因（Akr1b8 和 Rgs2）敲除的转录特征映射到与染色体 8 跨 eQTL 特征相关的 F2 网络模块，并且这些模块反过来与 F2 群体中的肥胖非常显著相关。总体而言，这项研究表明，如何将基因表达数据与网络框架中的 QTL 分析相结合，有助于阐明从老鼠到人类可以转化的疾病的分子驱动因素。
