Non-target screening and identification of the significant quality markers in the wild and cultivated Cordyceps sinensis using OPLS-DA and feature-based molecular networking

Cordyceps sinensis is a rare traditional Chinese herbal material. The cultivated cordyceps sinensis instead of the wild is a potential important development trend. Assessing the consistency of the wild and cultivated cordyceps sinensis is a very significant attention. The method of non-target screening and mining of the significant quality markers in the wild and cultivated cordyceps sinensis samples was established by liquid chromatography-quadrupole time-of-flight high resolution mass spectrometry (LC-Q-TOF-MS) combined with feature-based molecular networking (FBMN) and the model of orthogonal partial least square discriminant analysis model (OPLS-DA). Forty-seven training samples, including thirty-three wild cordyceps sinensis samples and fourteen cultivated cordyceps sinensis samples, were firstly used to build the OPLS-DA original model based on the 6827 feature m/z peaks. The 1144 feature m/z peaks of dimensionality reduction were built the OPLS-DA optimized model, which were acquired by the variable projected importance (VIP) of the OPLS-DA original model. Twenty nine significant markers were mined by using the S-plot of the OPLS-DA optimized model. Moreover, 17 of 29 significant markers were identified by the non-target screening of FBMN, including eight wild markers and nine cultivated markers. Finally, an overall correct rate of 95.5% (twenty two test samples) was obtained for classification of the wild and cultivated cordyceps sinensis samples based on the twenty nine significant markers. It is indicated that the significant quality markers of cordyceps sinensis could be mined and identified based on the non-target screening of FBMN coupled with OPLS-DA.