A robust EMT-related prognostic model for ccRCC integrating single-cell and bulk transcriptome data

Existing EMT-related prognostic models for clear cell renal cell carcinoma (ccRCC) are based solely on bulk transcriptomic data, lacking insights from single-cell RNA sequencing (scRNA-Seq). scRNA-seq and bulk transcriptomic data were utilized to identify EMT-related genes in ccRCC. A relevant prognostic model was constructed using univariate Cox regression analysis, 101 machine learning algorithms, and multivariate Cox regression analysis, which was further validated using three validation datasets: GSE167573, E-MTAB-1980, and ICGC. Ultimately, an EMT-related prognostic model consisting of AFM, CYS1, FAM171A1, GSTM3, FKBP10, MALL, RGS5, and TIMP1 was developed. This model effectively predicts the prognosis of ccRCC patients, and the risk score is associated with tumor mutation burden, immune cell infiltration, immune checkpoints, and different outcomes of immunotherapy. Simultaneously, AKT inhibitor VIII with strong correlations to the eight genes was identified. Molecular docking results showed high binding affinity, suggesting that the genes in this model could serve as potential targets for AKT inhibitor VIII. In summary, a more effective EMT-related prognostic model was constructed and AKT inhibitor VIII was identified as a potential targeted therapeutic agent.