Unveiling functionality and conducting two-sample mendelian randomization on WGCNA-identified oxidative stress-related hub genes in metabolic dysfunction-associated fatty liver disease

Abstract

Metabolic dysfunction-associated fatty liver disease (MAFLD) shows accelerated development under the impact of oxidative stress (OS). There is an imperative to identify OS-related biomarkers in MAFLD and explore their potential mechanistic insights. The objective of this study was to identify OS-related biomarkers in MAFLD and explore their potential mechanisms. DEG analysis was performed using GSE17470 and GSE24807 datasets. An investigative exploration utilizing WGCNA was executed to elucidate hub OS-related genes. The intersection of OS-related hub genes identified by WGCNA and DEGs was systematically employed for thorough analyses. A mendelian randomization (MR) study examined the causal effect of C-reactive protein (CRP) on MAFLD. 59 OS-related DEGs were identified in MAFLD. WGCNA revealed 100 OS-related hub genes in MAFLD. Sixteen OS-related genes have been delineated as critical components in MAFLD. Enrichment analyses, employing GO and KEGG pathways, revealed pathways enriched with these genes. Following PPI analyses, the highest-ranking ten hub genes demonstrating abnormal expression were determined. Ultimately, a two-sample MR analysis demonstrated a causal link between the hub gene CRP and the occurrence of MAFLD. In this study, we harnessed WGCNA to formulate a co-expression network and identified hub OS-related DEGs in MAFLD. Additionally, the hub gene CRP exhibited a significant correlation with the predisposition to MAFLD. These findings offer innovative perspectives on the applications of OS-associated genes in individuals afflicted with MAFLD.

Keywords: Gene co-expression network; MAFLD; Mendelian randomization; Oxidative stress; WGCNA.

Fig. 1

DEGs analysis reveals OS-related distinct profiles between MAFLD and normal groups. (A) Volcanic map illustrating OS-related DEGs in GSE17470. (B) Volcanic map illustrating OS-related DEGs in GSE24807. (C) Heat map illustrating OS-related DEGs of GSE17470. (D) Heat map illustrating OS-related DEGs of GSE24807. (E) The intersection of OS-related DEGs within the GSE17470 and GSE24807 datasets. OS: Oxidative stress; MAFLD: Metabolic dysfunction associated fatty liver disease; DEGs: Differentially expressed genes.

Fig. 2

MAFLD-associated gene modules identification in the GEO datasets through WGCNA. (A) Dendrogram representing clustering of all genes in GSE17470 based on topological overlap matrix, with each branch corresponding to a gene and different colored branches denoting co-expression modules. (B) Dendrogram representing clustering of all genes in GSE24807 based on topological overlap matrix, with each branch corresponding to a gene and different colored branches denoting co-expression modules. (C)Module-trait heatmap displaying the correlation between clustering gene module and MAFLD in the GSE17470 dataset, including correlation coefficients and p Values for each module. (D) Module-trait heatmap indicating the correlation between clustering gene module and MAFLD in the GSE24807 dataset, with corresponding correlation coefficients and p Values for each module. (E) Scatter plot showing the strongest positive correlation of module green with MAFLD in the GSE17470 dataset. (F) Scatter plot demonstrating the strongest positive correlation of module blue with MAFLD in the GSE24807 dataset. (G) The overlapping set of OS-associated genes, as identified through WGCNA, across the GSE17470 and GSE24807 datasets. OS: Oxidative stress; MAFLD: Metabolic dysfunction associated fatty liver disease; WGCNA: Weighted gene co-expression network analysis. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Unveiling candidate hub OS-related genes in MAFLD through DEGs and WGCNA analysis. (A) Venn diagram delineating the intersection of 16 candidate hub genes. (B) Bubble plot illustrating GO enrichment analysis of the candidate hub genes. (C) Circular plot showcasing GO enrichment analysis of the candidate hub genes. (D) Bubble plot presenting KEGG pathway analysis of the candidate hub genes. GO: Gene ontology; OS: Oxidative stress; MAFLD: Metabolic dysfunction associated fatty liver disease; WGCNA: Weighted gene co-expression network analysis. KEGG: Kyoto encyclopedia of genes and genomes.

Fig. 4

Depicts the establishment of the PPI network. (A) Depicts the PPI network featuring OS-related hub genes with significant overlap. (B) The extraction of pivotal genes within the interaction network was accomplished by employing the degree centrality algorithm. PPI: Protein-protein interaction; OS: Oxidative stress.

Fig. 5

The scatter plot distinctly portrays the significant causal influence of CRP on the risk of MAFLD in the MR study. (A) The significant causal influence of CRP (ebi-a-GCST90018730) on the risk of MAFLD (ebi-a-GCST90091033); (B) The significant causal influence of CRP (ebi-a-GCST90018730) on the risk of MAFLD (finn-b-NAFLD); (C) The significant causal influence of CRP (ukb-d-30710_raw) on the risk of MAFLD (ebi-a-GCST90091033); (D) The significant causal influence of CRP (ukb-d-30710_raw) on the risk of MAFLD (finn-b-NAFLD). CRP: C-reactive protein; MAFLD: Metabolic dysfunction associated fatty liver disease; MR: mendelian randomization.

Fig. 6

The forest plot delineates the causal impact of each SNP on the susceptibility to MAFLD in the MR study. (A) The causal impact of each SNP (ebi-a-GCST90018730) on the susceptibility to MAFLD (ebi-a-GCST90091033); (B) The causal impact of each SNP (ebi-a-GCST90018730) on the susceptibility to MAFLD (finn-b-NAFLD); (C) The causal impact of each SNP (ukb-d-30710_raw) on the susceptibility to MAFLD (ebi-a-GCST90091033); (D) The causal impact of each SNP (ukb-d-30710_raw) on the susceptibility to MAFLD (finn-b-NAFLD). MAFLD: Metabolic dysfunction associated fatty liver disease; SNPs: single-nucleotide polymorphisms; MR: mendelian randomization.

Fig. 7

Funnel plots were employed to visually represent the overarching heterogeneity of MR estimates regarding the impact of CRP on MAFLD. (A) The overarching heterogeneity of MR estimates regarding the impact of CRP (ebi-a-GCST90018730) on MAFLD (ebi-a-GCST90091033); (B) The overarching heterogeneity of MR estimates regarding the impact of CRP (ebi-a-GCST90018730) on MAFLD (finn-b-NAFLD); (C) The overarching heterogeneity of MR estimates regarding the impact of CRP (ukb-d-30710_raw) on MAFLD (ebi-a-GCST90091033); (D) The overarching heterogeneity of MR estimates regarding the impact of CRP (ukb-d-30710_raw) on MAFLD (finn-b-NAFLD). CRP: C-reactive protein; MAFLD: Metabolic dysfunction associated fatty liver disease; MR: mendelian randomization.

Fig. 8

Leave-one-out plot was applied to visualize the causal effect of CRP on the risk of MAFLD by leaving out one SNP in each iteration. (A) The causal effect of CRP (ebi-a-GCST90018730) on the risk of MAFLD (ebi-a-GCST90091033) by leaving out one SNP in each iteration; (B) The causal effect of CRP (ebi-a-GCST90018730) on the risk of MAFLD (finn-b-NAFLD) by leaving out one SNP in each iteration; (C) The causal effect of CRP (ukb-d-30710_raw) on the risk of MAFLD (ebi-a-GCST90091033) by leaving out one SNP in each iteration; (D) The causal effect of CRP (ukb-d-30710_raw) on the risk of MAFLD (finn-b-NAFLD) by leaving out one SNP in each iteration. CRP: C-reactive protein; MAFLD: Metabolic dysfunction associated fatty liver disease; MR: mendelian randomization; SNPs: single-nucleotide polymorphisms.