The ability of an electronic nose, based on a 6 metal oxide semiconductor chemical sensors array, to diagnose fungal contamination, to detect high fumonisin content and to predict fumonisin concentration was evaluated in vitro. Maize cultures were inoculated with Fusarium verticillioides, F. proliferatum, F. oxysporum, F. semitectum, F. solani and F. subglutinans. Sterilised maize cultures were used as reference. Fungal colonies and fumonisin content of the maize samples were used as covariates for statistical analyses and for electronic nose training. Univariate and multivariate exploratory data analysis showed that the electronic nose discriminated the inoculated maize culture samples according to their fumonisin content. Partial Least Square was also implemented to build a multivariate regression model based on EN signals for quantitative fumonisin prediction. The EN could correctly recognize high and low fumonisin content of maize cultures and provide a fair quantitative estimation. The validity of EN technology to perform a rapid screening of maize cultures in order to identify levels of fumonisin contamination below the acceptability threshold was established. (C) 2011 Elsevier Ltd. All rights reserved.
Electronic nose predicts high and low fumonisin contamination in maize cultures
Sberveglieri Giorgio
2011
Abstract
The ability of an electronic nose, based on a 6 metal oxide semiconductor chemical sensors array, to diagnose fungal contamination, to detect high fumonisin content and to predict fumonisin concentration was evaluated in vitro. Maize cultures were inoculated with Fusarium verticillioides, F. proliferatum, F. oxysporum, F. semitectum, F. solani and F. subglutinans. Sterilised maize cultures were used as reference. Fungal colonies and fumonisin content of the maize samples were used as covariates for statistical analyses and for electronic nose training. Univariate and multivariate exploratory data analysis showed that the electronic nose discriminated the inoculated maize culture samples according to their fumonisin content. Partial Least Square was also implemented to build a multivariate regression model based on EN signals for quantitative fumonisin prediction. The EN could correctly recognize high and low fumonisin content of maize cultures and provide a fair quantitative estimation. The validity of EN technology to perform a rapid screening of maize cultures in order to identify levels of fumonisin contamination below the acceptability threshold was established. (C) 2011 Elsevier Ltd. All rights reserved.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
