Functional investigation and two-sample Mendelian randomization study of Early gastric cancer (EGC) hub genes obtained by WGCNA analysis

Abstract

Abstract Objective: This study aimed to identify differentially expressed genes associated with early gastric cancer and analyze their potential functions through bioinformatics analysis. Additionally, we sought to validate the genetic causality of identified differential molecules related to early gastric cancer using Mendelian randomization. Methods: We analyzed the expression profiles of 19 gastritis specimens and 20 early gastric cancer specimens (high-grade intraepithelial neoplasia) from the GEO database using combined differential gene expression analysis and weighted gene co-expression network analysis (WGCNA). Hub genes most relevant to early gastric cancer were selected and subjected to functional enrichment and signaling pathway analysis using GO and KEGG enrichment. The diagnostic efficacy of the predictive model based on the top five ranked hub genes was evaluated using ROC curves. Finally, Mendelian randomization analysis was employed to assess genetic causality between hub genes and early gastric cancer. Results: A total of 755 hub genes were identified after intersecting the most relevant modular genes from WGCNA with genes exhibiting significant expression differences. GO and KEGG pathway enrichment analyses revealed significant differential expression of hub genes in responses to xenobiotic stimuli, hormone transport, apical cell parts, and oxidoreductase activity targeting the CH-OH group of donors, as well as involvement in pluripotency of stem cells, hepatocellular carcinoma, and axon guidance. The top five core genes—IL6, CLU, UGT2B15, NANOG, and NFE2L2—were analyzed using Cytoscape software. The column-linear graph prediction model demonstrated excellent performance in predicting the risk of early gastric cancer, as evidenced by ROC analysis. In the inverse variance weighting (IVW) method, we found that the core gene CLU was associated with an increased risk of early gastric cancer (OR = 1.157, 95% CI = 1.043-1.283, p = 0.0057). Conclusion: Our bioinformatics analysis identified the CLU gene as genetically causally associated with early gastric cancer, suggesting its potential as a diagnostic or therapeutic target for this condition.