ATR-FTIR spectroscopy combined with metabolomics to analyze the taste components of boletus bainiugan at different drying temperatures

Boletus bainiugan has a unique flavor profile, its quality is correlated with metabolites. Herein, ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) is utilized to characterize the free amino acid and organic acid of Boletus bainiugan at different drying temperatures. Attenuated total internal reflectance Fourier transform infrared (ATR-FTIR) spectroscopy is employed to identify Boletus bainiugan with various treatment and to predicted compounds. The metabolome includes 72 amino acids and 64 organic acids, wherein, 11 important taste components are analyzed the changes with drying temperatures. The residual convolutional neural network (ResNet) model achieves 100 % accuracy for Boletus bainiugan with distinct treatment. The partial least squares regression (PLSR) model accurately predicted the contents of 11 compounds with an optimal R2 P of 0.975 and a best residual predictive deviation (RPD) of 4.404. The ATR-FTIR spectroscopy coupled with metabolomics can be used as a good tool to estimate the taste enhancers of Boletus bainiugan.