OTA-grapes: a prototype model to predict ochratoxin A risk in grapes.

Ochratoxin A (OTA), a nephrotoxin and possible carcinogen fungal metabolite, can be produced by black Aspergilli in grapes and Aspergillus carbonarius is considered the main responsible for OTA contamination. A. niger and A. tubingensis were confirmed as dominant on berries, but they poorly contribute to OTA production. The European Commission set 2 µg/Kg as maximum OTA content allowed in wine, must and grape juice being wine the second most important source of OTA in the human diet, after cereals. In this context, it is important for stakeholders in the grape-wine chain to have an assessment of OTA risk related to the geographic area and year. Therefore, the aim of this study was to develop a mechanistic predictive model describing fungal infection and subsequently the risk of OTA contamination in grapes, along the growing season till to harvest. The model is focused on the key fungus A. carbonarius and uses weather data (air temperature, relative humidity and rain) as input. A relational diagram was drawn and data available in literature were used to develop mathematical functions for each step of the infection cycle (A. carbonarius sub-model), in relation to grape growth stages (grapes sub-model). In particular for sporulation, germination and infection, due to limited data available, only limit conditions for their occurrence were established. Fungal growth and OTA production were modelled at different temperature and water activity regimes using Bete and linear regression equation, respectively. Wide variability was noticed regarding OTA production by different strains, i.e. different combinations of optimal temperature and incubation time to optimize the production, apparently not related to the geographic origin of the strains. Therefore, two different functions were developed and combined to simulate the fungal community in vineyard. The A. carbonarius sub-model was run from setting to ripening, growth stages predicted by the grapes sub-model, that uses as well meteorological data as input. A prototype model is now available and the collection of a suitable set of vineyard data, including geo-referenced OTA contamination and related meteorological data is mandatory for the next step of model validation, crucial before the delivery of a predictive model as support for stakeholders in the grape chain.