Abstract
Genistein has anti-cancer effects, but its molecular targets in gastric adenocarcinoma (GA) are unclear. This study used single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics (ST) to explore genistein's "drug-gene-cell" interactions in GA. GA- and genistein-related target genes were retrieved and intersected with differentially expressed genes identified from bulk transcriptomic data. Machine learning screened candidates, and survival analysis assessed prognosis. Molecular docking with genistein validated key genes, with molecular dynamics assessing binding stability. HSD17B1, EZH2, CCNB1, CCNB2, CDKN2A, and IGFBP6 were identified as key candidate genes with prognostic value for GA. Specifically, samples in the IGFBP6 high-expression group were associated with higher survival probability, whereas the opposite trend was observed for the other five genes. In addition, HSD17B1 was genistein's main target in GA treatment, showing a strong binding affinity with genistein (binding energy of -8.1 kcal/mol). scRNA-seq analysis indicated that HSD17B1 was predominantly expressed in epithelial cells and was significantly involved during their malignant transformation (confirmed by ST). This study identified HSD17B1 as a critical target gene for genistein in GA treatment, emphasizing its roles in the malignant transformation of epithelial cells, thus providing a theoretical foundation for understanding the therapeutic mechanism of genistein in GA.