Integrated weighted gene co-expression network analysis and machine learning analysis identifies SEC14L5 as a potential biomarker for polycystic ovary syndrome

Abstract Background and aims To identify novel biomarkers and therapeutic targets for polycystic ovary syndrome (PCOS) using integrated bioinformatics approaches. Methods We integrated two microarray datasets (GSE34526, n = 10; GSE137684, n = 12) and validated findings in an independent RNA-seq dataset (GSE168404, n = 10). We employed weighted gene co-expression network analysis (WGCNA), functional enrichment analysis, and three machine learning algorithms to investigate differentially expressed genes (DEGs) and module genes. Results We retrieved gene expression datasets GSE34526 and GSE137684, utilizing the limma package to identify DEGs between PCOS and control subjects. WGCNA revealed 122 upregulated and 431 downregulated genes across 11 distinct modules, with the darkslateblue module containing 143 genes showing the highest Pearson correlation coefficient. Enrichment analyses indicated significant associations with pathways related to lipid metabolism, glucose metabolism, neutrophil regulation, and various immune functions. These findings were validated in an in vitro PCOS cell model. Conclusions Our study highlights SEC14L5 as a key differentially expressed gene in PCOS, providing a promising target for clinical research and treatment of PCOS patients.