Immune-related biomarkers for major depressive disorder identified via integrated bioinformatics and machine learning

Background: Major depressive disorder (MDD) is a leading cause of disability worldwide, yet its early and objective diagnosis remains challenging due to the lack of reliable biomarkers. Recent advances in high-throughput transcriptomics and machine learning provide new opportunities for systematic biomarker discovery. Methods: We integrated gene expression datasets from the Gene Expression Omnibus (GEO) to identify differentially expressed genes (DEGs) in MDD. Functional enrichment analyses were performed to explore biological relevance. To enhance robustness, two complementary machine learning algorithms – LASSO and SVM-RFE – were applied to screen candidate biomarkers. Diagnostic performance was assessed using receiver operating characteristic (ROC) curve analysis. Immune relevance was examined by CIBERSORT and validated in single-cell RNA sequencing (scRNA-seq) data. Finally, expression of hub genes was experimentally verified in a chronic unpredictable mild stress (CUMS) rat model. Results: A total of 122 DEGs were identified, primarily enriched in immune and inflammatory pathways. Four hub genes – DDIT4, DHRS9, FKBP5, and GPER – were consistently selected across machine learning approaches. These genes exhibited strong diagnostic accuracy (AUC values ranging from 0.82–0.91) and were predominantly expressed in immune cell populations. scRNA-seq further confirmed their upregulation in specific immune cell subtypes. Experimental validation showed significantly elevated expression of these genes in the prefrontal cortex of depressed rats. Conclusion: This study identifies DDIT4, DHRS9, FKBP5, and GPER as immune-related biomarkers with high diagnostic potential for MDD. By integrating bioinformatics, machine learning, and experimental validation, our work provides novel insights into the immune mechanisms underlying MDD and establishes a translational framework for precision diagnosis and personalised intervention. Integration of bioinformatics and machine learning enabled the identification of novel gene biomarkers for major depressive disorder (MDD), improving diagnostic prediction and potential therapeutic targeting.Immune cell infiltration analysis using multi-omics data revealed distinct immunological signatures in MDD, highlighting the crucial role of immune and inflammatory mechanisms in disease pathogenesis.Validation in the prefrontal cortex (PFC) emphasised its central role in MDD pathophysiology and provided a reliable brain region for biomarker discovery and translational application. Integration of bioinformatics and machine learning enabled the identification of novel gene biomarkers for major depressive disorder (MDD), improving diagnostic prediction and potential therapeutic targeting. Immune cell infiltration analysis using multi-omics data revealed distinct immunological signatures in MDD, highlighting the crucial role of immune and inflammatory mechanisms in disease pathogenesis. Validation in the prefrontal cortex (PFC) emphasised its central role in MDD pathophysiology and provided a reliable brain region for biomarker discovery and translational application.