Integrative multi‑cohort transcriptomics and network toxicology prioritize CYP1A2 as a hepatocyte‑enriched putative target of eugenol in hepatocellular carcinoma

Hepatocellular carcinoma (HCC) remains a leading cause of cancer mortality globally. Eugenol, a natural phenolic compound, shows therapeutic potential in various cancer models, yet its anti-HCC mechanisms remain largely unexplored. We conducted this computational multi-omics study by integrating multi-cohort transcriptomics, network-based target prediction, machine learning, molecular modeling, and single-cell RNA sequencing analysis to identify and prioritize potential key molecular targets of Eugenol in HCC. Transcriptomic analysis of six GEO datasets (716 samples) identified 449 differentially expressed genes (DEGs). Weighted gene co-expression network analysis (WGCNA) identified disease-associated co-expression modules containing 1,688 genes. Integration of DEGs, WGCNA results, and Eugenol-predicted targets (via SwissTargetPrediction, ChEMBL, and SEA databases) yielded 55 candidate genes. An ensemble machine learning framework (113 algorithm-feature selection combinations) with SHAP interpretability analysis identified eight key driver genes: CYP1A2, CXCL12, CYP3A4, CYP2C19, ALPL, AURKA, AURKB, and CYP17A1. CYP1A2 achieved the highest diagnostic utility (AUC = 0.902). Molecular docking and 100-ns molecular dynamics simulations suggested a structurally plausible interaction between eugenol and CYP1A2, with relatively stable conformational behavior observed during the simulated trajectory. Single-cell RNA-seq analysis (GSE146409) revealed hepatocyte-predominant CYP1A2 expression, approximately 3-fold higher than other liver cell types. Functional enrichment analysis highlighted enrichment in drug metabolism and catalytic pathways. Our findings prioritize CYP1A2 as a hepatocyte-enriched, HCC-associated candidate for potential interaction with Eugenol, providing a computational basis for future biochemical and functional validation.