Construction of the predictive model for idiopathic pulmonary fibrosis: Based on endoplasmic reticulum stress and cuproptosis related genes

Objective: Idiopathic pulmonary fibrosis (IPF) is a progressive, fatal lung disease with limited treatment options. This study aimed to construct a predictive model for IPF based on genes related to endoplasmic reticulum (ER) stress and cuproptosis, two pathophysiological processes implicated in IPF, to improve early diagnosis and understand underlying molecular mechanisms. Methods: Two public gene expression datasets (GSE33566 and GSE38958) were used to identify differentially expressed genes (DEGs) associated with ER stress and cuproptosis. Consensus clustering was applied to define molecular subtypes of IPF. Functional enrichment analysis was conducted via gene set variation analysis (GSVA). Weighted gene co-expression network analysis (WGCNA) was performed to identify key gene modules. Multiple machine learning models, including Random Forest (RF), were constructed and evaluated to develop a robust diagnostic signature. A nomogram was created to visualize risk prediction, and the relationship between model genes and pulmonary function (DLCO) was assessed. Results: A total of 537 DEGs related to ER stress and cuproptosis were identified in IPF. Two molecular subtypes of IPF showed distinct pathway enrichment and gene expression profiles. WGCNA revealed 56 hub genes, and the RF model based on five key genes (eukaryotic translation elongation factor 1 gamma (EEF1G), actin binding LIM protein 1 (ABLIM1), Kruppel like factor 12 (KLF12), zinc finger protein 573 (ZNF573), ras homolog family member H (RHOH)) achieved strong diagnostic performance (AUC = 0.829 in training; 0.700 in external validation). Of these, four genes correlated positively with DLCO, suggesting clinical relevance. Conclusion: This study demonstrates that integrating ER stress and cuproptosis gene signatures can effectively distinguish IPF patients. The findings offer novel insights into IPF pathogenesis and highlight potential biomarkers for early detection and personalized therapy.