Reduction strategies for farmers, i.e. in the field (pre-harvest), were tested for their efficacy: In field- and glasshouse experiments biopesticides and biofumigation techniques were tested for wheat and oats were tested in the UK and Norway; resistance of maize cultivars against Aspergillus flavus infection were tested in fields in Italy and Serbia; and an atoxigenic Aspergillus flavus strain was isolated and tested in fields in Serbia. These strategies for mycotoxin reduction in maize grown in Southern Europe showed very promising results: Using resistant plant cultivars resulted in up to 98% reduction of aflatoxin B1 at normal yields, while the use of an atoxigenic strain showed aflatoxin B1 reductions of up to 73%. Biocontrol strategies for wheat and oats in the UK and Norway were optimised. These efforts were complemented by developing European models that predict DON contamination in wheat, and fumonisins and aflatoxin contamination in maize, up to 4 days in advance. For biofuel production recombinant enzymes were added to bioethanol production: FB1 and ZEN could be simultaneously degraded up to 99% and 89%, respectively, in lab-scale tests. Formation of the degradation products hydrolyzed FB1 and hydrolyzed ZEN confirmed the biotransformation.
Catering to silo managers’ needs, sensors were developed to predict potential fungal growth – and thus mycotoxin contamination – in wheat and maize stored in European silos. Their functionality and underlying algorithms were tested and improved in pilot-scale silos at BARILLA. By including CO2 as an early-warning parameter, ZEN contamination in wheat and AFB1 contamination in maize could be forecasted 3-5 days earlier than using only temperature-sensitive sensors. These efforts were complemented by developing similar prediction models for peanuts stored in large scale silos in China.
Evaluating milling and baking techniques provided encouraging results for food producers and regulators: Different sorting, milling and micronization techniques were studied and combined to reduce DON contamination in the final product. The optimum balance of low DON at high fibre content was achieved by using medium-sized bran fraction combined with hammer milling and sieving sifter, resulting in a final product in which DON levels are well below the maximum limits and total dietary fibre is signifcantly higher than in standard wholewheat products. DON transformation during baking was assessed by using a C12/C13 labelling LC-HRMS metabolomics approach, pioneering a full mass balance of DON degradation during thermal processing : DON degraded by 6% in crackers, 5% in biscuits and 2% in bread, isoDON (1.3–3.9%) being the major degradation product. In vitro translation experiments indicate isoDON being less toxic than DON.
Within MyToolBox we also developed a novel sorting system for dried figs based on a combination of optical sorting technologies and algorithms. The prototype was demonstrated to end users in October 2019, and achieved an increased accuracy of up to 80 % under UV light.
Overall, the MyToolBox e-platform was launched officially on February 28, 2020: Besides including the real-time forecasting of mycotoxins in fields and silos, available information such as Good Agricultural Practices, sampling schemes, scenarios for cropping seasons, regulations, etc. were analysed, combined and transformed into easy-to-understand text to ensure that end-users all along the food and feed chain find suitable information – informed and tested particularly by focus-group meetings with potential end-users held in the Netherlands, UK, Serbia, Italy, Norway and China.
