Depression Diagnosis: Bioinformatics Uncovers Cellular Abnormalities in Female-Specific Genes With Ai-Driven Blood Biomarker-Based Prediction

With an estimated 264 million people affected by MDD globally, there is an urgent need for novel approaches to diagnosis and treatment. This study explored the complex gene expression profiles linked to Major Depressive Disorder (MDD) using state-of-the-art Next-Generation Sequencing (NGS), namely RNA-Seq. There is a need for accurate, non-invasive diagnostics, particularly through gender-specific biomarker research in female RNA-Seq data. To address this disparity, RNA-Seq data was analyzed using Machine Learning (ML) techniques to categorize female-specific MDD, identifying 10 critical biomarkers and utilizing Artificial Intelligence (AI) and bioinformatics to extract transcriptome data from publicly available datasets, while employing Generative Adversarial Networks (GANs) for data augmentation due to the limited availability of large RNA-Seq datasets. Gene set enrichment analysis (GSEA) and other bioinformatics analyses further demonstrated that genes differentially expressed in MDD patients were enriched in pathways related to the cell cycle, neutrophil degranulation, resolution of sister chromatid cohesion, and the formation of mitotic spindles by EML4 and NUDC. Ten important biomarkers associated with MDD were identified by an integrative bioinformatics investigation: IGHV3-64D, IGHV5-10-1, IFI27, LILRB5, PBK, SIGLEC1, IFI44L, CDCA5, and SLC4A1. Machine learning models based on these genes achieved $95 \%$ accuracy, highlighting potential of RNA-Seq and ML to enhance MDD diagnosis and reveal its molecular mechanisms.