EXPLORATION OF POTENTIAL TARGETS OF DIOSMIN IN THE TREATMENT OF MAJOR DEPRESSIVE DISORDER BASED ON MULTI-MACHINE LEARNING AND EXPERIMENTAL VALIDATION

Major Depressive Disorder (MDD) is a common psychiatric condition marked by persistent sadness and loss of interest, affecting life quality. This study seeks to find Dios-min-derived biomarkers for MDD using multi-machine learning, analyzing four gene expression datasets with 183 MDD and 111 control samples. Differentially expressed genes (DEGs) linked to Diosmin were identified and analyzed through weighted correlation network analysis (WGCNA) to find gene modules related to MDD. Machine learning techniques screened and validated potential biomarkers, creating a predictive model that effectively distinguishes MDD from controls. Key MDD signatures were compared for specificity and sensitivity, while CIBERSORT analysis examined immune cell infiltration in MDD. Drug sensitivity predictions and molecular docking assessed Diosmin's therapeutic potential. In an experimental validation, a lipopolysaccharide (LPS)-induced depression model was used, along with behavioral assays to evaluate Diosmin’s antidepressant effects, supported by immunofluorescence staining of the hippocampus. Our study identifies novel Diosmin-derived biomarkers and provides a foundation for developing targeted therapeutic strategies for MDD, highlighting the potential of integrating multi-machine learning for biomarker discovery.