Prediction and Analysis of Quality Markers for Hedysari Radix Using Ultra‐High‐Performance Liquid Chromatography‐Tandem Mass Spectrometry, Network Pharmacology, Molecular Dynamics, and Chemometrics

To establish a scientific and reliable quality evaluation method for Hedysari Radix (HR), this study investigated changes in chemical components before and after the "Rubbing strip" (RS) was processed based on Quality Marker (Q-Marker) theory. By integrating ultra-high-performance liquid chromatography-tandem mass spectrometry quantitative analysis, network pharmacology, molecular dynamics, and chemometrics to elucidate compositional variations and predict potential quality markers systematically. Results showed significantly increased contents of formononetin, calycosin, liquiritigenin, and γ-aminobutyric acid after the RS was processed, while the contents of ononin, calycosin-7-glucoside, and vanillic acid were reduced. These findings suggest that RS processing may facilitate the transformation of isoflavone glycosides into aglycones through mechanochemical effects. Network pharmacology analysis further identified 10 immune-related compounds, including medicarpin, calycosin-7-glucoside, calycosin, ononin, and formononetin, as potential quality markers. Using chemometric methods such as hierarchical cluster analysis, principal component analysis, and Orthogonal partial least-squares discriminant analysis, robust discrimination models distinguishing RS from "Non-rubbing strip" samples were successfully developed based on the identified Q-Marker. This study, integrating chemical profiling and bioinformatics approaches, identified and predicted potential Q-Markers for HR, providing theoretical and methodological support for establishing a robust and scientifically quality evaluation system for HR.