Identification of RAB36 and NINL as potential diagnostic biomarkers for gestational diabetes mellitus and polycystic ovary syndrome

Gestational diabetes mellitus (GDM) and polycystic ovary syndrome (PCOS) are common endocrine and metabolic disorders in women, with PCOS increasing the risk of GDM. However, shared molecular markers that could support early diagnosis and the mechanisms underlying them remain insufficiently defined. This study integrated public transcriptome datasets for GDM and PCOS to identify shared differentially expressed genes, which were then analyzed via machine learning, diagnostic model construction, pathway analysis, regulatory network analysis, drug prediction, molecular docking, and reverse transcription (RT)-qPCR-based validation in clinical blood samples. Among the 15 shared candidate genes identified, RAB36 and ninein-like (NINL) were selected as potential diagnostic biomarkers. Nomogram and receiver operating characteristic (ROC) curve analyses indicated that the two-gene models have diagnostic value for both GDM and PCOS, with area under the curve (AUC) values of 0.71 and 0.98, respectively. Functional analyses suggested that RAB36 and NINL may be associated with pathways related to immune regulation, cell cycle-related processes, metabolic signaling, and neutrophil extracellular trap formation. Drug prediction and molecular docking further identified cinnamaldehyde and valproic acid as potential compounds targeting both biomarkers. RT-qPCR revealed increased RAB36 expression and decreased NINL expression in clinical blood samples. This study identifies potential biomarkers shared between GDM and PCOS and mechanistic clues for early diagnosis and personalized disease management.