Identification of key programmed cell death genes in chronic atrophic gastritis

Using the training set (GSE27411) and validation set (GSE229899) from the GEO database, along with PCD and metabolism-related genes obtained from literature and the KEGG database, data analysis was performed using GSVA, limma, WGCNA, Lasso regression, and SVM-RFE.Results Five PCD types (Parthanatos, Alkaliptosis, Cuproptosis, Ferroptosis, Pyroptosis) showed significant differential expression in atrophic gastritis. Through WGCNA and differential gene analysis, 10 candidate genes were identified, and further analysis using Lasso regression and SVM-RFE revealed 5 key genes (CASP1, CASP6, SLC40A1, GLS, CA9).

These genes demonstrated good diagnostic performance (AUC > 0.7) in both the training and validation sets and were significantly associated with the immune and metabolic microenvironment. Additionally, using DGIdb and mirNet, 47 potential therapeutic drugs and 24 transcription factors were predicted, and molecular docking validated the binding activity between key genes and drugs.Conclusion This study elucidates the mechanisms of PCD-related genes in atrophic gastritis and provides a theoretical basis for developing potential therapeutic targets.

Future research should expand sample sizes and conduct clinical validation to advance the application of related therapeutic strategies.