Integrating scRNA-seq and machine learning identifies MNAT1 as a therapeutic target in OSCC.

Background: Oral squamous cell carcinoma, with high global incidence and mortality, requires improved early intervention strategies. Ubiquitination - a critical post-translational modification - has been strongly implicated in tumorigenesis, with particularly significant roles in T-cell regulation. We developed a T Cell-Related ubiquitination risk model that enhances prognostic prediction and immunotherapy response assessment, offering a framework for personalized OSCC manageme. Method: T cell-Related Ubiquitination genes were identified based on scRNA-seq analysis, and key genes were selected using WGCNA and LASSO algorithms to construct a prognostic model. Spearman correlation analysis revealed significant associations between riskScore and immune infiltration levels, checkpoint molecule expression, and MMR activity. Pseudotemporal trajectory and cell-cell communication analyses delineated dynamic gene expression patterns driving OSCC progression. Functional validation through colony formation and Transwell assays confirmed the tumor-suppressive effects of key model genes. Results: Given the high correlation between T cell-Related Ubiquitination genes and the prognosis of OSCC patients, a prognostic model based on patient scRNA-seq data was constructed and validated. The RiskScore derived from our model correlated significantly with expression levels of MMR genes, abundance of immune checkpoint proteins, and immunotherapy response. Cell-cell communication analysis further elucidated epithelial-macrophage crosstalk via MIF and IFN-II signaling, suggesting microenvironment-driven progression mechanisms. Conclusions: Collectively, integrating T cell-Related Ubiquitination genes through advanced computational analyses, we established a robust prognostic model for OSCC and identified MNAT1 as a promoter of malignant progression, highlighting its therapeutic potential.