Integrative multi-omics and machine-learning approaches uncover a novel metabolic-related signature associated with cancer-associated fibroblasts in gastric cancer development.
Journal:
Computers in biology and medicine
Published Date:
Jun 20, 2025
Abstract
Gastric cancer (GC) ranks as the fifth most commonly diagnosed malignancy and the fourth leading cause of cancer-related mortality worldwide. The integration of machine learning in the analysis of GC metabolomics data for biomarker identification remains underexplored, presenting substantial opportunities for further investigation. This study integrated various data types, including microarray, bulk RNA-seq, and single-cell RNA sequencing (scRNA-seq). GSE15459 and GSE26253 were used as external validation cohorts to evaluate model performance. GC's metabolic-related signature (MRS) expression profiles in GC were analyzed using scRNA-seq data from the GSE183904 dataset. Notably, metabolism-related genes exhibited significantly higher enrichment in fibroblasts compared to other cell types, suggesting a potential role of fibroblasts in metabolic dysfunction associated with cancer progression. From this, the analysis revealed that only the gene signature associated with a cancer-associated fibroblast (CAF) cluster, annotated as CAF_6, was significantly correlated with poorer overall survival. This indicated that CAF_6 had a potential pro-tumorigenic role, enriched for TNF-alpha signaling via NF-kB and epithelial-mesenchymal transition pathways. Additionally, a combined machine learning strategy was employed to develop a metabolism-associated signature, demonstrating a strong association with prognosis in TCGA-STAD patients (p-value <0.01). A nomogram was constructed to predict clinical outcomes in STAD patients by incorporating the MRS risk score alongside clinical parameters such as disease stage, age, and sex. In conclusion, our study emphasizes the molecular and functional heterogeneity of CAFs in GC and introduces a novel MRS with strong prognostic relevance. These insights provide a foundation for future investigations aimed at refining risk stratification and advancing targeted therapeutic approaches for GC.
