Potential use of machine learning methods in assessment of Fusarium culmorum and F. proliferatum growth and mycotoxin production in treatments with antifungal agents

Abstract The use of Fusarium-controlling fungicides is necessary to limit crop loss. Little is known about the effect of commercial antifungal formulations at sub-lethal doses, and their interaction with abiotic factors, on Fusarium culmorum and F. proliferatum development and on zearalenone and fumonisin biosynthesis, respectively. In the present study different treatments based on sulfur, trifloxystrobin and demethylation inhibitor fungicides (cyproconazole, tebuconazole and prothioconazole) under different environmental conditions, in Maize Extract Medium (MEM), are assayed in vitro. Then, several machine learning methods (neural networks, random forest and extreme gradient boosted trees) have been applied and compared for the first time for modeling growth rate of F. culmorum and F. proliferatum and zearalenone and fumonisin production, respectively. The most effective antifungal treatment was prothioconazole, 250 g/L + tebuconazole, 150 g/L. Effective doses of this formulation for reduction or total fungal growth inhibition ranged as follows ED50 0.49–1.70, ED90 2.57–6.02 and ED100 4.0–8.0 μg/mL, depending on the species, water activity and temperature. Overall, the growth rate and mycotoxin levels in cultures decreased when doses increased. However, some treatments in combination with certain aw and temperature values significantly induced toxin production. The extreme gradient boosted tree was the machine learning model able to predict growth rate and mycotoxin production with minimum error and maximum R2 value.