A New Quality Control Strategy for Mori Ramulus Adulteration in Astragali Radix: A Qualitative-Quantitative Coupled Model Based on Non-Targeted Mass Spectrometry Digital Fingerprints and the ElasticNet Algorithm

Abstract

Astragali Radix (AR) is widely applied in the pharmaceutical and food fields, but it is highly susceptible to malicious adulteration with the visually highly similar Mori Ramulus (MR) during market circulation. To address these issues, This study constructed a qualitative–quantitative coupled adulteration analysis model based on a nontargeted UHPLC-QTOF-MS method by integrating digital fingerprints of AR and MR, ion mapping ratio (IMR) screening, and regression algorithms. Under the experimental conditions in this study, the qualitative screening model established based on IMR could stably detect MR adulteration as low as 5%. Combined with the quantitative prediction model constructed by ElasticNet algorithm (R2 = 0.9969, RMSE = 0.0195, MAE = 0.0130), two suspicious adulterated samples were collaboratively screened from 16 batches of AR blind samples (ABS), with predicted MR adulteration ratios of 42.48% and 77.43%, respectively. In conclusion, the digital fingerprints established via nontargeted UHPLC-QTOF-MS coupled with IMR and ElasticNet model can provide a feasible technical support for adulteration screening and the quantitative estimation of the adulteration ratio of AR with MR.