3D chromatin architecture-related genes orchestrate LUAD evolution and therapy resistance: insights from integrative machine learning and spatial single-cell mapping - PubMed
3 hours ago
- #3D chromatin architecture
- #machine learning
- #LUAD
- Identified 21 3D chromatin structure-related differentially expressed genes (3DDEGs) in LUAD tumors vs. normal tissues.
- Integrated machine learning across multiple GEO datasets pinpointed SMC3, RAD21, FOXM1, MYBL2, and MTA3 as key genes influencing early recurrence, Sorafenib response, and tumor stage progression.
- High-risk early-stage LUAD patients showed co-activated glycolysis/oxidative phosphorylation and increased myeloid-derived suppressor cell (MDSC) infiltration.
- Single-cell and spatial transcriptomics confirmed enrichment of these genes in tumor proliferating cells and epithelial hotspot regions.
- Virtual perturbation analysis revealed the core genes converge to regulate cell cycle, DNA replication, and mismatch repair pathways in tumor cells.
- Immunohistochemistry validated higher SMC3 protein expression in tumor tissues compared to adjacent normal tissues.
- The study underscores the role of 3D chromatin architecture in LUAD evolution and therapy resistance, offering insights for precision medicine.