Exploration of ferroptosis-related biomarkers and regulatory mechanisms in two diseases (sarcopenia and idiopathic pulmonary fibrosis) based on transcriptome and machine learning.

BACKGROUND

Sarcopenia (SP) and idiopathic pulmonary fibrosis (IPF) are debilitating diseases, but share ferroptosis-related biomarkers and their regulatory mechanisms in both remain unclear, hindering targeted research and therapy.

METHODS

In this study, ferroptosis-related genes (FeRGs), as well as training sets and test sets related to SP and IPF, were obtained from public databases. After the identification of disease-related differentially expressed genes (DEGs), DEGs shared by SP and IPF were obtained through intersection (genes with consistent DEG change trends across each disease) and union analyses. These shared DEGs were then intersected with FeRGs to obtain candidate genes. Machine learning algorithms and the Wilcoxon rank-sum test were used to confirm the biomarkers, and nomograms were constructed and evaluated. Meanwhile, in-depth studies such as functional enrichment, immune infiltration, drug prediction, and molecular docking were conducted on the biomarkers.

RESULTS

Initially, a total of 26 candidate genes were identified. After screening, DDIT4 and MGST1 were identified as ferroptosis-related biomarkers shared by the two diseases, and both were significantly upregulated in the SP and IPF groups. The area under the curve values of the constructed nomograms were 0.91 (for SP) and 0.79 (for IPF), respectively. The p-values of the Hosmer-Lemeshow test for the calibration curves (0.856 and 0.205) were both >0.05, and the DCA showed good performance. The two biomarkers showed differences in enriched pathways. A total of 5 differential immune cells shared by the two diseases were identified, and the biomarkers participated in immune regulation by targeting specific immune cells. Good binding activity (Vina Score < -5.0) was observed between Trichostatin A and both biomarkers.

CONCLUSION

DDIT4 and MGST1 were shared ferroptosis-related biomarkers for SP and IPF. Nomograms enabled reliable prediction, and trichostatin A was a potential targeted drug. These findings supported mechanistic research and targeted therapy for both diseases.