Integrating Metabolomics and Machine Learning to Analyze Chemical Markers and Ecological Regulatory Mechanisms of Geographical Differentiation in Thesium chinense Turcz

• DOI: 10.20944/preprints202505.1441.v1
• Publication Date: 2025-05-22T00:49:13Z
• Authors: Cong Wang; Ke Che; Guanglei Zhang; Hao Yu; Junsong Wang

The relationship between medicinal efficacy and geographical environment in Thesium chinense Turcz., a traditional Chinese herb, remains systematically unexplored. This study integrates metabolomics, machine learning, ecological factor analysis to elucidate the variation patterns regulatory mechanisms of secondary metabolites from Anhui, Henan, Shanxi provinces. Using UHPLC-Q-TOF MS coupled with PCA/PLS-DA multivariate analysis, we identified 43 marker compounds (primarily flavonoids alkaloids). Random forest LASSO regression algorithms determined core markers for each production area: Anhui (4 markers), Henan (6 (3 markers). Metabolic pathway enrichment revealed these exert pharmacological effects through neuroactive ligand-receptor interaction PI3K-Akt signaling pathways. Redundancy demonstrated samples exhibited significantly higher antioxidant activity (DPPH hydroxyl radical scavenging rates) than other regions, strongly correlating stable low-temperature environments (annual mean temperature) precipitation patterns. establishes first geo-specific molecular system chinense, providing scientific basis quality control geo-authentic herbs environmental adaptation research.