Systems genetics analysis of cardiovascular traits in a mouse intercross: Integration of expression data, clinical traits and functional information.

摘要

Metabolic syndrome represents a significant risk factor for diabetes and cardiovascular diseases, the leading causes of death in developed countries. To better understand the molecular and genetic factors associate with metabolic disorder, we developed and applied integrative genetic approaches to identity genes predisposing common and complex diseases in genetically segregating populations. Integrative genetics approach is designed to understand the interaction between gene expression, genotyping and phenotypic variations. The benefit of this approach is to provide a plausible hypothesis of molecular mechanisms for disease-associated variants. The characterization of transgenic mice over-expressing USF1, a disease gene of inheritable dyslipidmia, validated the functional role of this gene as a causal gene of metabolic traits in vivo. The application of gene network to analyze global gene expression revealed the involvement of immune responses and insulin signaling in the progress of USF1-caused trait variation. To further understand molecular events underlying metabolic disorder, we developed a regulatory network algorithm to suggest causal regulators for gene expression and their corresponding traits, termed PhenoNet. PhenoNet identifies regulators by translating function roles of SNPs measured in a population. We identified liver gene modules significantly associate with lipid and cardiovascular diseases in a hyperlipidemic mouse population. Perturbation of regulator genes in mouse primary hepatocytes validated the importance of regulator for module gene expression. Application of weighted gene co-expression network method to transcript data of a previously defined mitochondrial protein consortium revealed the functions and genetic variants associating with these genes. The construction of a two-tier gene co-expression network provided information about genes co-regulated with proteins associated with mitochondria in liver and adipose, a algorithm named as MAGEN. Taken together, our integrative genetics approaches implicate functional role and molecular network of genes underling common and complex traits for metabolic disorder. Further validation of prediction made with integrative genetics approaches in animal models will further reveal the molecular mechanisms of metabolic syndrome, diabetes and cardiovascular disease.