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.