Weighted gene co-expression network analysis identifies biomarkers in glycerol kinase deficient mice.

Symptomatic glycerol kinase deficiency (GKD) is associated with episodic metabolic and central nervous system deterioration. We report here the first application of weighted gene co-expression network analysis (WGCNA) to investigate a knockout (KO) murine model of a human genetic disease. WGCNA identified networks and key hub transcripts from liver mRNA of glycerol kinase (Gyk) KO and wild-type (WT) mice. Day of life 1 (dol1) samples from KO mice contained a network module enriched for organic acid metabolism before Gyk KO mice develop organic acidemia and die on dol3-4. Furthermore, the module containing Gyk was enriched with apoptotic genes. We used causal testing to elucidate the causal relationships between intramodular hub genes Acot, Psat and Plk3. Important causal relationships are confirmed in cell cultures. We provide evidence that GK may have an apoptotic moonlighting role that is lost in GKD. This first application of WGCNA to mouse knockout data provides insights into the molecular mechanisms of GKD pathogenesis. The resulting systems-genetic gene screening method identifies candidate biomarkers for GKD.