Multi-omics integration and machine learning reveal biomarker networks and therapeutic targets in Alzheimer's disease

Alzheimer's disease (AD) is a progressive neurodegenerative disorder with limited early diagnostic options and no curative therapy. This study presents a multi-region transcriptomic and systems biology framework for the identification of robust biomarker genes and pharmacologically actionable targets in AD. Differential gene expression profiles from the entorhinal cortex, hippocampus, and frontal cortex were analyzed using three supervised machine learning algorithms: LASSO regression, Random Forest, and SVM-RFE to prioritize predictive biomarkers. Consensus genes demonstrated region-specific differential expression and moderate diagnostic accuracy (AUC up to 0.70). Functional enrichment revealed their roles in synaptic transmission, translational regulation, lysosomal acidification, and extracellular matrix remodeling, hallmarks of AD pathology. Protein–protein interaction networks and hub gene analyses further underscore the role of translational regulators, such as EIF3C and FAU. Transcription factor mapping (e.g., MEF2C, ZBTB18), miRNA targeting (e.g., miR-107, miR-195–5p), and drug–gene interaction analysis identified GABBR2 and COL5A2 as translationally relevant targets linked to approved or investigational drugs. This integrative study proposes a reproducible pipeline combining machine learning, regulatory network modeling, and pharmacogenomic mining to inform biomarker-driven drug repurposing in Alzheimer's disease.