Background: Polycystic ovary syndrome (PCOS) is a common disease that often results in miscarriage among females. The study aimed to elucidate the relationship between chromatin regulators and PCOS to explore potential biomarkers for PCOS.

Methods: Differential genes between PCOS patients and healthy controls were screened based on GSE10946 dataset. Pivotal genes were screened using LASSO (Least Absolute Shrinkage and Selection Operator), XGBoost (eXtreme Gradient Boosting) and random forest, and crossover genes of these three machine learning methods were further screened using a Venn diagram. The predictive potential of these key genes was evaluated using receiver operating characteristic (ROC) curve analysis, and functional enrichment analysis was conducted to elucidate the associated signaling pathways. To investigate their roles in PCOS and their effects on the immune microenvironment, immunity profiles between PCOS patients and healthy controls were compared; the correlations between key genes and immune cell populations were also examined, and then the associations of these genes with potential drug candidates were explored.

Results: GADD45A and TAF5 were identified as critical genes associated with PCOS, exhibiting significantly decreased expression levels in PCOS patients compared to healthy controls (both p < 0.05). Both genes showed strong predictive performance. Functional enrichment revealed their involvement in pathways such as cholesterol homeostasis and E2F targets. Immune infiltration analysis indicated distinct differences in immune cell composition between the PCOS and control groups, with gene expression levels closely correlating with specific immune cell subsets (all p < 0.05).

Conclusion: GADD45A and TAF5 have high predictive ability for PCOS and have the potential to be new biomarkers and therapeutic targets for PCOS.