Construction of a prognostic model and multidimensional analysis of hepatocellular carcinoma based on...

This study aims to construct a prognostic model for hepatocellular carcinoma (HCC) based on palmitoylation-related genes and explore its molecular mechanisms through multi-dimensional analyses.

The research integrated single-cell transcriptome data (GSE189903) with bulk transcriptome data (TCGA-LIHC, GEO datasets), focusing on palmitoylation-related genes in HCC epithelial cells. The scAB deconvolution algorithm was used to analyze the association between epithelial cell subsets and patient survival, and hdWGCNA was combined to construct a gene co-expression network. Through differential expression analysis, univariate Cox regression, and LASSO penalized regression, 7 key genes (SERPINE1, FMO3, ALDH2, CPS1, SLCO1B1, ACAT1, ACADS) were identified to build a prognostic risk model.

Validation results showed that the model could effectively distinguish the survival prognosis of high-risk and low-risk patients (AUC values for 1/3/5 years in the TCGA cohort were 0.676, 0.656, and 0.642, respectively; those in the GSE14520 validation set were 0.702, 0.658, and 0.654, respectively), and the risk score was an independent prognostic factor. Further analyses revealed that the risk score was associated with tumor staging, immune cell infiltration (e.g., T cells, monocytes), response to immunotherapy, and drug sensitivity. Functional enrichment analysis indicated that the high-risk group was enriched in cell cycle regulation and oncogenic signaling pathways, while the low-risk group was related to metabolic pathways.

This study is the first to analyze the regulatory network of palmitoylation in HCC epithelial cells by combining single-cell and bulk transcriptomes, providing new molecular targets and methodological references for HCC prognosis evaluation and precision therapy.