Aims: To develop and validate a logistic regression model to predict the growth and ochratoxin A (OTA) production boundaries of two Aspergillus carbonarius isolates on a synthetic grape juice medium as a function of temperature and water activity (a(w)).
Methods and results: A full factorial design was followed between the factors considered. The a(w) levels assayed were 0.850, 0.880, 0.900, 0.920, 0.940, 0.960, 0.980 and the incubation temperatures were 10, 15, 20, 25, 30, 35 and 40 degrees C. Growth and OTA production responses were evaluated for a period of 25 days. Regarding growth boundaries, the degree of agreement between predictions and observations was >99% concordant for both isolates. The erroneously predicted growth cases were 3.4-4.1% false-positives and 0.7-1.4% false-negatives. No growth was observed at 10 degrees C and 40 degrees C for all a(w) levels assayed, with the exception of 0.980 a(w)/40 degrees C, where weak growth was observed. Similarly, OTA production was correctly predicted with a concordance rate >98% for the two isolates with 0.7-1.4% accounting for false-positives and 2.0-2.7% false-negatives. No OTA production was detected at 10 degrees C or 40 degrees C regardless of a(w), and at 0.850 a(w) at all incubation temperatures. With respect to time, the OTA production boundary shifted to lower temperatures (15-20 degrees C) as opposed to the growth boundary that shifted to higher temperature levels (25-30 degrees C). Using two literature datasets for growth and OTA production of A. carbonarius on the same growth medium, the logistic model gave one false-positive and three false-negative predictions out of 68 growth cases and 13 false-positive predictions out of 45 OTA production cases.
Conclusions: The results of this study suggest that the logistic regression model can be successfully used to predict growth and OTA production interfaces for A. carbonarius in relation to temperature and a(w).
Significance and impact of the study: The proposed modelling approach helps the understanding of fungal-food ecosystem relations and it could be employed in risk analysis implementation plans to predict the risk of contamination of grapes and grape products by A. carbonarius.