Background:Studies have demonstrated the prognostic potential of Epithelial-mesenchymal transition (EMT)-related genes (ERGs) and ferroptosis-related genes (FRGs) for Lung adenocarcinoma (LUAD) patients, but their correlation and underlying molecular mechanisms remain unknown. Thus, this investigation employs transcriptomics, Mendelian randomization (MR) analysis and single-cell data to explore the relationship and prognosis of EMT and ferroptosis with LUAD.
Methods:Transcriptomic data from training and validation sets, along with single-cell datasets, were utilized. Prognostic-related differentially expressed (DE)-FRGs and DE-ERGs were identified, and a risk model was constructed using MR analysis and machine learning techniques. Single-cell data were analyzed to characterize model genes’ expression and function.
Results:Totally, 40 DE-FRGs and 222 DE-ERGs related to prognostic were identified in TCGA-LUAD. And there were 7 candidate genes by MR analysis. A total of 6 model genes (TFAM, ALOX15, GATA1, ALAD, PDCD4 and NDRG2) were selected to construct a risk model and the model was verified by GSE11969. Meanwhile, stage, N, T, and risk score were screened out as independent prognostic factors via univariate and multivariate cox regression and scores were counted to build a nomogram. Then, MR analysis results revealed that 6 model genes were significantly causally associated with LUAD. Meanwhile, model genes were mainly enriched in NK cells, T cells, Endothelial cell clusters, etc. by single-cell analysis.
Conclusion:TFAM, ALOX15, GATA1, ALAD, PDCD4 and NDRG2 played crucial roles in LUAD prognosis, providing refresh directions for the treatment of LUAD.