Identification of Glycosylation-Related Biomarkers in COPD and IPF Through Integrated Machine Learning and WGCNA Analysis.

Objective: This research aimed to explore key glycosylation-related genes (signature genes) and associated molecular mechanism on chronic obstructive pulmonary disease (COPD) and idiopathic pulmonary fibrosis (IPF), which further providing new perspectives for disease prognosis and diagnose. Patients and Methods: The gene expression profiles were obtained from the public GEO database. The glycosylation-related genes were identified based co-DEGs from COPD vs normal samples and IPF vs normal samples, module genes by weighted gene co-expression network analysis (WGCNA), as well as glycosylation genes from database. Signature genes were screened using machine learning methods, followed by immune infiltration, function analysis, drug-gene and transcriptional regulatory network analysis. Finally, validation analysis based on tissue samples from COPD/IPF patients were performed to test the expression of signature genes. Results: A total of 35 differentially expressed glycosylation-related genes for both COPD and IPF were explored. By three kinds of machine learning analyses, totally three signature genes including SULF1, ST8SIA1 and FCN3 were explored. In COPD, the AUC values for FCN3, ST8SIA1, and SULF1 were 0.643, 0.722, and 0.719, respectively; while in IPF, they were 0.955, 0.792, and 0.943, respectively. Immune infiltration and GSEA analysis showed that signature genes were dramatically correlated with activated B cell and extracellular matrix (ECM)-associated functions ( Conclusion: We identified SULF1, ST8SIA1 and FCN3 as shared glycosylation-related biomarkers in COPD and IPF. These genes bridge fibrosis, inflammation, and immune dysregulation, offering potential diagnostic and therapeutic targets.