Rapid metabolic profiling and authentication of Cordyceps using ambient ionization mass spectrometry and machine learning.

Cordyceps sinensis, a symbiotic organism formed between a fungus and an insect, is celebrated for its substantial medicinal benefits and economic significance in traditional Chinese medicine. However, the market for Cordyceps sinensis is rife with counterfeits, where numerous types of Cordyceps frequently pose as the genuine species, leading to financial losses for consumers. Here, we developed an ambient ionization mass spectrometry for the metabolic analysis of four kinds of Cordyceps. We tentatively identified a total of 81 metabolites, revealing significant differences between wild-type Cordyceps sinensis and its counterfeit counterparts. The heterogeneous distribution of metabolites was also examined. Notably, ergothioneine, an antioxidant, and its precursor hercynine were found to be more abundant in the stroma compared to other sections. Then, a neural network was employed to distinguish between different Cordyceps, achieving an average classification accuracy of 90.3% in blind tests. We demonstrate the potential for on-site detection of Cordyceps using a handheld nano-electrospray ionization source in conjunction with a miniature mass spectrometer, yielding mass spectral profiles comparable to those obtained with a benchtop system.