Abstract: This study employed near-infrared spectroscopy (NIRS) to acquire spectral data from 160Ganodermaspp. mycelium samples. The β-1,3-glucan content was measured using the aniline blue fluorescence method as the reference value. A predictive model for β-1,3-glucan content inGanodermaspp. mycelium was established based on partial least squares regression (PLSR). After excluding outliers, 150 valid samples were retained. Thirteen characteristic wavelengths were extracted using the competitive adaptive reweighted sampling (CARS) algorithm. Five machine learning algorithms were used for modeling,including partial least squares regression (PLSR), back propagation neural network (BPNN), extreme learning machine (ELM), support vector machine (SVM) and random forest (RF). The PLSR model achieved the highest accuracy (R2=0.946, RMSEP=0.545, RPD=4.303). Predictions and validation of β-1,3-glucan content in 20Ganodermaspp. mycelium samples using this model showed satisfactory accuracy and robustness (R2=0.8425, RMSEP=0.8025, RPD=2.4121, all exceeding 2.4). This study developed a practical β-1,3-glucan assay method, enabling rapid and non-destructive quantitative detection of β-1,3-glucan inGanodermaspp. mycelium samples.

Key words: Ganodermaspp. mycelium, β-1,3-glucan, near infrared spectral, regression prediction, PLSR