Predictive models for assessing the risk of Fusarium pseudograminearum mycotoxin contamination in post-harvest wheat with multi-parameter integrated sensors

•

Water activity plays a significant role in affecting CO₂ and mycotoxin levels.

•

Models were developed to predict contamination with ZEN and DON in stored wheat.

•

These models + multi-parameter integrated sensors for real-time mycotoxin monitoring.

Reliable prediction of the risk of mycotoxin contamination in post-harvest wheat will aid in improvement of the quality and safety. To establish the relationship between Fusarium pseudograminearum mycotoxins and CO₂ production, changes in their respective concentrations were monitored for the artificial contamination of wheat under different values of water activities (0.84 a_w, 0.92 a_w, and 0.97 a_w) and temperatures (20 ℃, 25 ℃, and 30 ℃). Water activity played a significant role in all these processes. CO₂ concentration together with moisture content and temperature were used as the main parameters to establish DON and ZEN contamination prediction models. The prediction accuracy for DON was 98.15 % (R² = 0.990) and 90.74 % for ZEN (R² = 0.982). These models were combined with T/RH/MC/CO₂ multi-parameter integrated sensors to form an early warning system, which offers a great prospect to minimise the risk of DON/ZEN contamination in post-harvest wheat.