Machine Learning-Derived Diverse Regulated Cell Death Patterns for Therapeutic Target Identification in Glaucoma.

Purpose: Glaucoma is the leading cause of irreversible vision loss worldwide. We aimed to uncover the molecular mechanisms and regulatory networks of hub genes in human glaucoma to identify promising targets for detection and treatment. Methods: We obtained GSE758, GSE2378, and GSE9944 datasets from the Gene Expression Omnibus database. The list of genes linked to regulated cell death (RCD) was obtained from a previous study. RCD-related differentially expressed genes (DEGs) were identified in patients with glaucoma and controls. Weighted Gene Co-Expression Network Analysis (WGCNA) and machine learning algorithms were used to identify hub genes. Gene set enrichment analysis (GSEA) was used to explore signaling pathways enriched by hub genes, and molecular docking analysis was performed to identify the gene-drug network of hub genes for potential treatment. Immunofluorescence was used to reveal the expression levels of hub genes in glaucomatous mice and controls. Results: This study identified 358 RCD-related DEGs that distinguished healthy individuals from glaucoma patients and underscored the pivotal involvement of the immune response in human glaucoma pathogenesis. We systematically identified 33 hub genes, including PLEC, DLGAP4, Glycosylphosphatidylinositol (GPI), etc. that demonstrated significant diagnostic or treatment potential for glaucoma. The cytoskeletal regulator PLEC has emerged as a promising candidate gene associated with glaucomatous neurodegeneration with possible acting drugs. Conclusion: We constructed a machine-learning-driven analytical framework based on diverse RCD patterns to refine molecular subtypes and druggable genes. These findings may provide novel targets for glaucoma detection and treatment.