Fresh Portulaca oleracea L. exhibits stronger anti-gastric cancer activity than dried purslane: Insights from integrated metabolomics and machine learning

Purslane (Portulaca oleracea L.) is a valuable edible plant with chemopreventive properties; however, the impact of post-harvest processing, specifically drying, on its bioactive profile and anti-gastric cancer efficacy remains poorly understood. This study employed an integrated foodomics strategy combining metabolomics and machine learning to systematically elucidate the mechanisms underlying the effects of fresh purslane (FP) and dried purslane (DP). Metabolomic analysis indicates that the drying process causes degradation of key phenolic and flavonoid compounds in fresh purslane, thereby reducing its antioxidant activity and its ability to inhibit the proliferation of HGC-27 gastric cancer cells. To explore the molecular basis underlying these differences, a multi-model machine learning approach prioritized PDZ-binding kinase (PBK) as a critical candidate target. Subsequent molecular docking analysis of the differential metabolites identified baicalin as a candidate bioactive constituent associated with PBK-related regulation. Experimental validation demonstrated that baicalin inhibited HGC-27 cell viability and induced apoptosis with G2/M phase arrest. Baicalin treatment reduced PBK mRNA expression and promoted PBK protein degradation, suggesting a suppressive effect on PBK expression and stability. Collectively, these findings show that fresh purslane exerts stronger anti-gastric cancer activity than dried purslane and support the use of integrated metabolomics and machine learning to uncover molecular clues underlying this difference.