| Congresso Brasileiro de Microbiologia 2023 | Resumo: 199-1 | ||||
Resumo:The economic losses and potential health risk, due to contamination by mycotoxins, resulting by the presence of the complex of species ,Fusarium graminearum in Brazil’s barley require severe concern to the agricultural and food sectors. Previous studies focused on Brazilian barley contamination by Fusarium species showed a high occurrence of F. graminearum s.s., F. meridionale, F. armineacum, F. poae, and F. cortaderie. Considering the variety of Fusarium species and the diversity of metabolites produced, the identification and characterization of this group is relevant. However, the accuracy in identifying them is a problem commonly encountered by researchers, as many are morphologically and biologically critical. The next generation sequencing (NGS) information about the whole genome is facilitating the inference and distinction of phylogenetic species, however, the high cost associated with the need for specific training hinders its wide use. In this context, alternative methods to toxigenic fungi characterization have been tested, as vibrational spectroscopy techniques. This study aimed to test the suitability of UV-Vis spectroscopy data coupled to data analysis tools to classify F. graminearum species. It was selected F. poae, F. armineacum and F. graminearum, three species previously isolated from Brazilian barley samples. These species were cultivated in V8 culture medium, at 25°C for 7 days, the mycelium was removed from the petri dish with a scalpel and placed into a microtube, where 1mL of deionized water was added, it was homogenized and followed by centrifugation, the supernatant was recovered and 200 µL was placed into a microplate where the Uv-Vis spectra data, from 200 nm – 1000 nm with resolution of 1 nm, was measured. The spectra data was mean centered and the principal component analysis (PCA) and classification model Partial Least Square – Discriminant Analysis (PLS-DA) determined. Three principal components (PCs) were set, with the cumulative variance of 89.95%. The scores plot based on PC1 versus PC2 revealed a great tendency group for each species, indicating that the UV-Vis spectral information had a great potential to classify Fusarium species. The PLS-DA model was determined with three latent variables, a cross-validation based on venetian blinds algorithm was performed, using the most probable prediction, the final results were sensitivity and sensibility were 100%. The results obtained in this study showed a great potential of Uv-Vis spectroscopy to identification of Fusarium species, further investigations will be conducted in order of increase the number of species and to test the discriminant ability of UV-Vis technique. Palavras-chave: Chemometrics, Barley, fungi Agência de fomento:FAPESP (Process number: 2021/14060-1) | |||||