Xiaorong Zhang, Yufan Zeng, Jiehan Zhang, Weiye Duan, Xinhua Ao, Huizhu Wang, Shuai Chen
Polysaccharides from Chroogomphus rutilus (CRPs) exhibit promising bioactivities; however, conventional alkaline extraction (CAE) often results in low yields and structural disruption. In this study, a support vector regression (SVR)–optimized ultrasonic-assisted alkaline extraction (UAAE) process was developed to maximize yield while preserving conformational integrity. Compared with the quadratic response surface methodology model, SVR showed superior predictive accuracy and robustness, with the testing R 2 increasing from 0.8751 to 0.9027 and RMSE reduced by 11.7%. SVR also exhibited narrower residual dispersion and improved stability in the high-yield region (>17%), confirming its suitability for nonlinear, multivariable bioprocess optimization. Under the optimized conditions—ultrasonic temperature 56.5 °C, ultrasonic time 35 min, soaking time 138 min, liquid–solid ratio 26 mL/g, NaOH concentration 0.55 mol/L, and ultrasonic power 325 W—the extraction yield reached 20.09 ± 0.10%, representing a 53.7% increase compared with CAE. High-performance gel permeation chromatography revealed two dominant molecular weight (Mw) fractions for UAAE-derived CRP at 7.54 × 10 5 and 1.32 × 10 4 Da, whereas CAE-derived products exhibited a trimodal distribution dominated by low-Mw species at 1.43 × 10 3 Da. Spectroscopic, microscopic, and thermal analyses demonstrated that UAAE more effectively preserved ordered helical conformations, and improved structural stability compared with CAE. CRPs obtained under optimized conditions also showed enhanced antioxidant activity, including ABTS• + and hydroxyl radical scavenging capacities. The integrating acoustic cavitation with alkaline treatment and SVR-based modeling provides an efficient, data-driven strategy for sustainable production of high-quality fungal polysaccharides with preserved bioactive conformations.