Identifying diagnostic biomarkers for glaucoma based on transcriptome combined with Mendelian randomization

Glaucoma poses a significant global health challenge, yet reliable biomarkers for its diagnosis and treatment remain scarce. This study employed Mendelian randomization (MR) and bioinformatics approaches to identify potential biomarkers for glaucoma. Using the GSE9944 dataset, differentially expressed genes (DEGs) were identified and analyzed through protein-protein interaction (PPI) networks and functional enrichment. MR analysis selected DEGs for further evaluation using support vector machine-recursive feature elimination (SVM-RFE), with genes exhibiting high differential expression and an area under the curve (AUC) > 0.7 considered as candidate biomarkers. Among 836 DEGs, the PPI network revealed complex interactions, and functional enrichment highlighted significant involvement of the PI3K-AKT and MAPK signaling pathways. MR analysis linked 113 DEGs to glaucoma, with 57 genes showing consistent expression trends. SVM-RFE identified six signature genes, among which ATP6V0D1 and FAM89B emerged as robust biomarkers (AUC > 0.7). Molecular regulatory network analysis and drug prediction analysis further revealed potential mechanisms and compounds targeting these biomarkers, providing new therapeutic avenues for glaucoma. Experimental validation confirmed that ATP6V0D1 and FAM89B were significantly downregulated under both mechanical and swelling stress conditions, with concurrent suppression of the PI3K/AKT pathway. In conclusion, ATP6V0D1 and FAM89B are promising biomarkers for glaucoma, offering potential applications in diagnosis, treatment, and advancing the understanding of glaucoma pathogenesis.