Identification of Gene Modules and Hub Genes Involved in Mastitis Development Using a Systems Biology Approach

Abstract

Objective: Mastitis is defined as the inflammation of the mammary gland, which impact directly on the production performance and welfare of dairy cattle. Since, mastitis is a multifactorial complex disease and the molecular pathways underlying this disorder have not been clearly understood yet, a system biology approach was used in this study to a better understanding of the molecular mechanisms behind mastitis.

Methods: Publicly available RNA-Seq data containing samples from milk of five infected and five healthy Holstein cows at five time points were retrieved. Gene Co-expression network analysis (WGCNA) approach and functional enrichment analysis were then applied with the aim to find the non-preserved module of genes that their connectivity were altered under infected condition. Hub genes were identified in the non-preserved modules and were subjected to protein-protein interactions (PPI) network construction.

Results: Among the 25 modules identified, eight modules were non-preserved and were also biologically associated with inflammation, immune response and mastitis development. Interestingly most of the hub genes in the eight modules were also densely connected in the PPI network. Of the hub genes, 250 genes were hubs in both co-expression and PPI networks and most of them were reported to play important roles in immune response or inflammatory pathways. The blue module was highly enriched in inflammatory responses and STAT1 was suggested to play an important role in mastitis development by regulating the immune related genes in this module. Moreover, a set of highly connected genes were identified such as BIRC3, PSMA6, FYN, F11R, NFKBIZ, NFKBIA, GRO1, PHB, CD3E, IL16, GSN, SOCS2, HCK, VAV1 and TLR6, which have been established to be critical for mastitis pathogenesis.

Conclusion: This study improved the understanding of the mechanisms underlying bovine mastitis and suggested eight non-preserved modules along with several most important genes with promising potential in etiology of mastitis.

Keywords: RNA-seq; bovine; hub genes; protein-protein interaction; weighted gene co-expression network.

FIGURE 1

The used pipeline for construction of the co-expression gene network.

FIGURE 2

Clustering dendrogram of genes based on a dissimilarity measure (1-TOM), which was then used to group genes into 25 modules in healthy samples. The branches correspond to modules of highly interconnected groups of genes. The height (y-axis) indicates the co-expression distance and the x-axis corresponds to genes. Colors represent the 25 different modules along with gray indicating genes that could not be assigned to any module.

FIGURE 3

The medianRank (left scatter plot) and Zsummary (right scatter plot) statistics of the module preservation. The medianRank and Zsummary of the modules close to zero indicate the high and low degree of module preservation, respectively. A negative Zsummary value indicates the modules’ disruption.

FIGURE 4

GO analysis results of highly and semi-preserved modules. Owing to the large number of significant GO terms (biological process) in the green and pink modules, only the top 20 significant terms are displayed. Size and color of points represent -Log2 of FDR and number of genes associated with each term, respectively.

FIGURE 5

GO analysis results of non-preserved modules. Owing to the large number of significant GO terms (biological process) in blue, brown, red, magenta, black, and purple modules, only the top 10 significant terms are displayed. Size and color of points represent -Log2 of FDR and number of genes associated with each term, respectively. The fold enrichment measure is not shown for better visualization.

FIGURE 6

PPI network based on the hub genes of the blue module. The larger circles indicate highly connected genes and the orange circle represents the only TF among the highly connected genes.