Identification of a ubiquitin related gene signature for predicting prognosis and immune infiltration in breast cancer based on single cell and bulk RNA sequencing analysis

Background Ubiquitination drives breast cancer (BC) progression, yet the clinical value of ubiquitin-related genes (UbRGs) remains underexplored. This study aimed to develop a UbRG-based prognostic model and investigate its associations with the tumor immune microenvironment (TIME) and therapeutic responses. Methods Transcriptomic and clinical data from the TCGA-BRCA cohort were analyzed to identify differentially expressed UbRGs. A ubiquitin-related risk score (URS) was constructed using LASSO and Cox regression. Single-cell RNA sequencing (scRNA-seq) data (GSE161529) were integrated via the Scissor algorithm to identify risk-associated cell populations. Furthermore, weighted gene co-expression network analysis (WGCNA) combined with machine learning approaches was applied to identify a core hub gene, which was subsequently validated using immunohistochemistry (IHC) in a clinical cohort. Results A six-gene URS independently stratified patients into high- and low-risk groups with significantly different overall survival. High-risk tumors exhibited an immunosuppressive TIME, featuring decreased lymphocyte infiltration and M2 macrophage enrichment, which was corroborated at the single-cell level. Conversely, low-risk patients demonstrated predicted sensitivity to targeted therapies, including CDK4/6 and PARP inhibitors. Through multi-omics and machine learning integration, SORBS1 was identified as the critical hub gene. IHC validation confirmed that SORBS1 protein expression was significantly downregulated in BC tissues and correlated with improved survival. Conclusions The scRNA-seq and machine learning-integrated URS provides robust prognostic stratification and therapeutic guidance in BC. Our findings link ubiquitination to immune exclusion and establish SORBS1 as a clinically validated, protective biomarker.