Graph-Based Multi-Omics Integration Reveals Prognostic Histone Modification Reader Genes and Candidate Drug Targets in Colorectal Cancer.

BACKGROUND: Colorectal cancer (CRC) is driven by genetic alterations, epigenetic dysregulation and tumor microenvironment remodeling. Histone modification reader proteins serve as key epigenetic regulators of anti-tumor immunity, yet their synergistic immune networks, combined prognostic roles and immune subtype heterogeneity remain poorly understood. METHODS: Here, we integrated multi-omics data and graph attention networks (GAT) to systematically screen prognostic-associated histone reader genes. We then conducted analyses using the immunoassay pipeline and ultimately identified hub immune-related genes validated in independent external cohorts. Additional analyses, including single-cell RNA sequencing (scRNA-seq), molecular docking and multiple in silico functional assays, were performed based on retrospective public datasets. RESULTS: Four core genes (CUL7, GPC1, NFYA, SLC25A5) exhibited robust prognostic performance and strong correlations with anti-tumor immunity. Both core and auxiliary genes participate in critical metabolic and immune pathways. Candidate drugs present differential binding affinity for their encoded proteins, with sapitinib designated as a promising agent. CONCLUSIONS: This work constructs an epigenetic immune regulatory network and a four-gene signature, offering promising biomarkers and actionable therapeutic targets for precision immunotherapy against CRC.