PDK4 and RGMA as Biomarkers for Gastric Cancer Diagnosis and Prognostication: Transcriptomic and Mendelian Randomization Analyses

Background: Gastric cancer has a poor prognosis and remains a major public health challenge. Therefore, identifying reliable biomarkers is essential to enhance early detection and improve patient survival. Methods: We integrated gene expression data from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) to identify markedly dysregulated genes, and performed Mendelian randomization analyses using MR-Base to evaluate potential causal relationships. Colocalization analysis was subsequently conducted to refine candidate loci. Functional enrichment analyses, including Gene Set Enrichment Analysis (GSEA) and Gene Set Variation Analysis (GSVA), were performed to explore the underlying biological processes, followed by construction of a regulatory network and development of a prognostic model. We further established and validated a risk model for predicting overall survival. Single-cell RNA sequencing data (GSE163558) were analyzed to characterize gene expression across specific cell types. Finally, immunohistochemistry (IHC) was used to verify protein-level differences between gastric cancer tissues and adjacent normal tissues. Results: Pyruvate dehydrogenase kinase 4 (PDK4) and repulsive guidance molecule A (RGMA) were identified as biomarkers. These genes play active roles in key biological processes, including cellular transformation and angiogenesis. By integrating PDK4 and RGMA expression with clinical parameters, we constructed a prognostic model that accurately predicted 3- and 5-year survival outcomes in the TCGA-STAD cohort. Single-cell analysis further revealed cell–type–specific expression patterns of these genes. In addition, immunohistochemical assays demonstrated higher protein levels in tumor tissues compared with adjacent normal tissues. Conclusions: This comprehensive study suggests that PDK4 and RGMA are promising biomarkers that may facilitate the early detection of gastric cancer and improve prognostic assessment. These findings provide new insights into the disease's pathogenesis and may further inform future clinical decision-making.