Identification of Fusarium sambucinum species complex by surface-enhanced Raman spectroscopy and XGBoost algorithm

Rapid and reliable identification of Fusarium fungi is crucial, due to their role in food spoilage and potential toxicity. Traditional identification methods are often time-consuming and resource-intensive. This study explores the use of surface-enhanced Raman spectroscopy (SERS) to identify four species from the Fusarium sambucinum species complex isolated from barley. SERS spectra from 60 samples was acquired using gold nanoparticles for signal enhancement and the eXtreme Gradient Boosting (XGBoost) algorithm was applied for classification. The method achieved 100 % precision, recall, accuracy, and F1-score, thereby demonstrating excellent performance. Regarding the chemical interpretability, key spectral features at 495, 546, 764, 1228, 1274, and 1605 cm^-1 were revealed by XGBoost and correlated to the differences in chemical composition of fungi; particularly related to chitin, metabolites, and protein content. Therefore, SERS and XGBoost have great potential to classify a wide variety of fungi and other microorganisms.