Identification of potential biological processes and key genes in diabetes-related stroke through weighted gene co-expression network analysis

Abstract

Background: Type 2 diabetes mellitus (T2DM) is an established risk factor for acute ischemic stroke (AIS). Although there are reports on the correlation of diabetes and stroke, data on its pathogenesis is limited. This study aimed to explore the underlying biological mechanisms and promising intervention targets of diabetes-related stroke.

Methods: Diabetes-related datasets (GSE38642 and GSE44035) and stroke-related datasets (GSE16561 and GSE22255) were obtained from the Gene Expression omnibus (GEO) database. The key modules for stroke and diabetes were identified by weight gene co-expression network analysis (WGCNA). Gene Ontology (GO) and Kyoto Encyclopedia of Genes Genomes (KEGG) analyses were employed in the key module. Genes in stroke- and diabetes-related key modules were intersected to obtain common genes for T2DM-related stroke. In order to discover the key genes in T2DM-related stroke, the Cytoscape and protein-protein interaction (PPI) network were constructed. The key genes were functionally annotated in the Reactome database.

Results: By intersecting the diabetes- and stroke-related crucial modules, 24 common genes for T2DM-related stroke were identified. Metascape showed that neutrophil extracellular trap formation was primarily enriched. The hub gene was granulin precursor (GRN), which had the highest connectivity among the common genes. In addition, functional enrichment analysis indicated that GRN was involved in neutrophil degranulation, thus regulating neutrophil extracellular trap formation.

Conclusions: This study firstly revealed that neutrophil extracellular trap formation may represent the common biological processes of diabetes and stroke, and GRN may be potential intervention targets for T2DM-related stroke.

Keywords: Bioinformatics; Diabetes; Neutrophil extracellular trap formation (NETs); Stroke; Weight gene co-expression network analysis (WGCNA).

Fig. 1

The research flowchart of data preparation and analysis. SVA: surrogate variable analysis; WGCNA: weighted gene co-expression network analysis; GSEA: Gene set enrichment analysis; GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; PPI: protein–protein interaction

Fig. 2

Batch effects considered in analysis. A, B the distribution of stroke samples before elimination of batch effect. C, D the distribution of stroke samples after eliminating the batch effect

Fig. 3

Batch effects considered in analysis. A, B the distribution of diabetes samples before elimination of batch effect. C, D the distribution of diabetes samples after eliminating the batch effect

Fig. 4

Identification and pathway analyses of differentially expressed genes (DEGs). A Heatmap of DEGs in diabetes related datasets; B Volcano plots showing the differential genes in diabetes related dataset. C Heatmap of DEGs in stroke related datasets; D Volcano plots showing the differential genes in stroke related dataset. E, F Ridgeline plot showing KEGG pathways enrichment in diabetes (E) and stroke (F). G, H Gene set enrichment analysis (GSEA) plots showing the most enriched gene sets of all detected genes in the diabetes (G) and stroke (H) samples

Fig. 5

Construction of weighted co-expression network for stroke-related datasets. A Network topology analysis of different soft threshold power. B Dendrograms of genes acquired by mean linkage hierarchical clustering. C The relationship between trait and modules. Correlation coefficients and corresponding P value are listed for each module. D The genes in the salmon module were significantly correlated with stroke. E Heatmap depicts the Topological Overlap Matrix (TOM) of genes selected for weighted co-expression network analysis. F Heatmaps and hierarchical cluster dendrograms of clinical traits and module eigengene. G, H The GO and KEGG pathways in key modules

Fig. 6

Construction of weighted co-expression network for diabetes-related datasets. A The value of scale independence on the left and the value of mean connectivity on the right. B The relationship between trait and modules. Correlation coefficients and corresponding P value are listed for each module. C The cluster dendrogram of all gene were grouped into different modules. D The genes in the steelblue module were significantly correlated with diabetes. E Network heatmap of all genes (a color change from red to yellow indicates a high degree of overlap between modules). F Heatmaps and hierarchical cluster dendrograms of clinical traits and module eigengene. G, H The GO and KEGG pathways in key modules

Fig. 7

Screening of the common genes between stroke and diabetes. A The common genes screened by intersecting the diabetes and stroke related modules. B The co-enriched biological processes by metascape

Fig. 8

Screening the hub genes and the related biological process analysis. A a PPI network of common genes constructed by Cytoscape software. B, C the role of hub gene GRN regulatory pathways in inflammatory response (B) and neutrophil degranulation (C). The schematic art pieces were provided by the Reactome Database (www.reactome.org) under Creative Commons Attribution 4.0 License