Rapid discrimination and quantification of chemotypes in Perillae folium using FT-NIR spectroscopy and GC-MS combined with chemometrics.

Perillae Folium (PF) is a well-known food and herb containing different chemotypes, which affect its quality. Herein, a method was proposed to classify and quantify PF chemotypes using gas chromatography-mass spectrometry (GC-MS) and Fourier transform-near infrared spectroscopy (FT-NIR). GC-MS results revealed that PF contains several chemotypes, including perilla ketone (PK) type, -asarone (PP-as) type, and dillapiole (PP-dm) type, with the PK type being the predominant chemotype. Based on FT-NIR data, different chemotypes were accurately classified. The random forest algorithm achieved >90 % accuracy in chemotype classification. Furthermore, the main components of perilla ketone and isoegomaketone in PF were successfully quantified using partial least squares regression models, with prediction to deviation values of 3.76 and 2.59, respectively. This method provides valuable insights and references for the quality supervision of PF and other foods.