Machine learning identifies lipid-associated genes and constructs diagnostic and prognostic models for idiopathic pulmonary fibrosis.

OBJECTIVE

Emerging evidence suggests a potential relationship between lipid metabolism and idiopathic pulmonary fibrosis (IPF). This study aimed to identify lipid-related genes implicated in IPF pathogenesis.

METHODS

Lipid-associated genes were retrieved from the GeneCards database and analyzed using unsupervised consensus clustering to classify IPF samples. Weighted gene co-expression network analysis (WGCNA) was performed on the identified clusters to determine core modules and hub genes associated with IPF. Machine learning algorithms were applied to these hub genes to construct diagnostic and prognostic models, which were validated across multiple datasets. Single-cell sequencing was used to investigate the distribution of potential pathogenic genes, and their functional roles were further validated through cellular experiments.

RESULTS

Two distinct clusters were identified, showing significant differences in lung function parameters and fibrosis-related gene expression. WGCNA revealed that the blue module was strongly associated with IPF and served as the core module. Genes from this module were used to construct diagnostic and prognostic models, which demonstrated strong predictive performance across multiple validation datasets. Single-cell sequencing revealed that KLF4 was highly expressed in lung epithelial cells. Functional assays indicated that knockdown of KLF4 did not affect the proliferation of human alveolar type II epithelial cells but significantly enhanced their migratory capacity, thereby promoting the fibrotic process.

CONCLUSION

This study successfully constructed lipid-related diagnostic and prognostic models for IPF and identified KLF4 as a potential causative gene. These findings provide a foundation for further exploration of lipid metabolism in IPF pathogenesis and potential therapeutic strategies targeting KLF4.