Analysis of characteristic aroma compounds in Gastrodia elata f. glauca S. Chow and identification of fine-grained sampling points based on molecular docking and machine learning

Gastrodia elata f. glauca S. Chow (WTM) is a both a traditional food and a traditional Chinese medicine plant. The study aimed to reveal its flavor variations between fine-grained sampling points and formation mechanism though Gas chromatograohy-mass spectrometry (GC-MS), Attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy, molecular docking, and machine learning. A total of 538 volatiles was detected, with 106 exhibited relative odor activity values (ROAVs) greater than 1. Molecular docking results indicated that the olfactory receptor binds effectively with 106 components, 31 of which were key flavor compounds. Hydrogen bonds and hydrophobic interactions were the main interaction forces. The temperature, precipitation and soil factors affect WTM aromas' formation and distribution. The radial basis function neural network (RBFNN) model had 100% accuracy on the training set and test set. This study provides a theoretical foundation and novel approach for flavor mechanism, quality influential factors and sampling point traceability of WTM.