An integrative in silico study using network toxicology and multi-omics reveals HSPA8 and MAPK3 as potential mediators linking barbecue-derived carcinogens to hepatocellular carcinoma

This in silico study aimed to explore the potential molecular mechanisms through which barbecue-derived toxicants may contribute to hepatocellular carcinoma (HCC) and to identify potential diagnostic biomarkers and therapeutic targets. By integrating network toxicology, multi-algorithm machine learning, weighted gene co-expression network analysis, and multi-omics analyses (transcriptomics, spatial transcriptomics, single-cell RNA-seq), this study constructed a combined analytical pipeline. Four barbecue toxicants (benzo[a]pyrene, 2-amino-1-methyl-6-phenylimidazopyridine, acrylamide, and nitrosamines) were assessed using ADMET and ProTox tools. Their potential targets were intersected with HCC-associated genes, followed by functional enrichment, protein-protein interaction network analysis, molecular docking, and virtual knockout simulations. Unimodal computational approaches showed limited mechanistic insight, whereas the integrated multi-omics strategy achieved optimal predictive performance through heterogeneous data synergy, identifying 279 overlapping targets and converging on six hub genes. Among these, HSPA8 and MAPK3 emerged as independent prognostic factors, both significantly upregulated in HCC tissues and correlating with poor survival. Molecular docking predicted favorable binding affinities to these targets. Immune profiling, single-cell, and spatial analyses supported tumor-dominant expression patterns. Virtual knockout simulations suggested that perturbation of HSPA8 or MAPK3 may dysregulates networks central to antigen presentation and T cell activation. The integrated multi-omics and computational fusion strategy could potentially mitigate the limitations of unimodal approaches, providing a framework with predictive accuracy and mechanistic interpretability for generating hypotheses on barbecue-related HCC pathogenesis and for individualized diagnosis and treatment.