Elucidating the murine brain transcriptional network in a segregating mouse population to identify core functional modules for obesity and diabetes

Abstract

Complex biological systems are best modeled as highly modular, fluid systems exhibiting a plasticity that allows them to adapt to a vast array of changing conditions. Here we highlight several novel network-based approaches to elucidate genetic networks underlying complex traits. These integrative genomic approaches combine large-scale genotypic and gene expression results in segregating mouse populations to reconstruct reliable genetic networks underlying complex traits such as disease or drug response. We apply these novel approaches to one of the most extensive surveys of gene expression studies ever undertaken in whole brain in a segregating mouse population. More than 23,000 genes were monitored in whole brain samples from more than 300 mice derived from an F2 intercross population and genotyped at over 1200 SNP markers uniformly spread over the entire genome. We explore the topological properties of the brain transcriptional network and highlight different approaches to inferring causal associations among genes by integrating genotypic and expression data. We demonstrate the utility of these approaches by identifying and experimentally validating brain gene expression traits predicted to respond to a strong expression quantitative trait locus (eQTL) for the pituitary tumor-transforming 1 gene (Pttg1) that coincides with the physical location of this gene (a cis eQTL). We identify core functional modules making up the brain transcriptional network in mice that are coherent for core biological processes associated with metabolic disease traits including obesity and diabetes. Keywords: genetics, gene expression, integration genomics, QTL, networks, systems biology. J. Neurochem. Complex traits like common human diseases and drug response involve complex interactions among genes and between genes and environment, and also among genetic, genomic, metabolomic, proteomic, and signaling networks. Whereas in the past the reductionist approach to elucidating complex systems was necessary for understanding fundamental building blocks of these systems and because tools did not exist to take on a more holistic approach to dissecting such systems, the explosion of large-scale, high throughput technologies in the biological sciences has motivated a rapid paradigm shift away from reductionism in favor of a systems level view of biology (Hartwell et al. 1999;Kirschner 2005;Schadt et al. 2005a). Recent studies have begun to characterize topological properties of biological networks in mammalian systems, focused primarily on protein interaction and

Systematic genetic and genomic analysis of cytochrome P450 enzyme activities in human liver

“…This provided a basis for the identification of key functional modules within the networks that contribute to variability of traits of interest. We have previously described the characterization of transcriptional coexpression networks based on brain, adipose, and liver tissues in human and mouse (Gargalovic et al 2006;Ghazalpour et al 2006;Horvath et al 2006;Lum et al 2006;Chen et al 2008;Emilsson et al 2008). Building on this approach, we constructed a coexpression network based on the human liver tissue data to identify gene modules.…”

“…A number of studies have previously demonstrated that coexpression networks are both scale-free and modular (Ghazalpour et al 2006;Lum et al 2006), thus highlighting functional components of the network that are often associated with specific biological processes. Therefore, to identify modules composed of highly interconnected expression traits within the coexpression network, we examined the topological overlap matrix (TOM) (Ravasz et al 2002) associated with this network.…”

“…In this study, we focused on the liver since it represents the primary location of activity for the majority of P450s, and attempted to address these gaps via integration of genetics, gene expression, and enzyme activity measurements in a human cohort. In addition, we employed a weighted gene coexpression network approach that has been used extensively in yeast, mouse, and human to uncover biologically meaningful gene modules and explore novel P450 functions/pathways and coregulation structure among P450s (Zhang and Horvath 2005;Gargalovic et al 2006;Ghazalpour et al 2006;Horvath et al 2006;Lum et al 2006;Chen et al 2008;Emilsson et al 2008;Schadt et al 2008;Zhu et al 2008a,b). Furthermore, networks constructed based on Bayesian inference were used as an alternative approach to delineate potential regulatory mechanisms.…”

Thematic review series: Systems Biology Approaches to Metabolic and Cardiovascular Disorders. Reverse engineering gene networks to identify key drivers of complex disease phenotypes

“…One of the modules highlighted in the Fig. 2A topological overlap map, generated from liver samples in a previously described cross (31,32), is enriched for genes associated with lipid and cholesterol The left panel represents a topological overlap map of the liver tissue from female mice in the BXH cross (24, 31, 32) constructed using previously described methods (56). The plot represents 5,000 of the most highly connected genes in the liver tissue of the BXH cross, with red and blue indicating positive and negative correlation, respectively, between the corresponding genes, and white indicating absence of correlation at some prespecified correlation threshold.…”

Expression Quantitative Trait Loci Mapping Identifies New Genetic Models of Glutathione S-Transferase Variation

“…This methodology can also identify regulators of transcriptional pathways, or expression quantitative trait loci (eQTL). eQTL analysis has been used to correlate genome-wide gene expression data with genetic variation to identify master-regulator transcription factors and their downstream targets in the brain and, in a different study, a novel component of the oxidative phosphorylation pathway (Lum et al, 2006;Wu et al, 2008).…”