Identification and characterization of gene networks and key genes related to the high-yield production of milk in high-yield cows using meta-analysis of microarray data

Abstract

Milk yield is most important economic trait in dairy cows and understanding molecular basis and components involved in high-yield production is one of crucial steps to develop and select new breeds. In this study, we used combination of two statistical methods based on the p-value and effect sized to meta-analysis three datasets followed with construction of weighted gene co-expression network based on the expression matrix of differentially expressed genes identified in meta-analysis to detect major gene modules and hub genes. Based on the FDR cut-off<0.05 and Log₂ fold change>2 and < 0.5, we identified 1028 differentially expressed genes that were shared between the Fisher and REM method and were consistent across datasets. Molecular function analysis showed that upregulated differentially expressed genes mostly enriched to ion binding, small molecule binding, and identical protein binding while downregulated genes were enriched to catalytic activity (Bonferroni test; threshold of 0.05). Weighted gene co-expression network analysis identified three major modules associated with fatty acid metabolism, PPAR signaling pathway, insulin resistance, terpenoid backbone biosynthesis, and steroid biosynthesis. A total of 12 hub genes (one downregulated and 11 upregulated) identified from protein-protein interaction network of modules. This study could identify new differentially expressed genes related to lactation processes in high-yield-cows. Moreover, we could reveal some gene modules and hub genes in each module which are biologically more meaningful.