Prediction of deoxynivalenol levels in Fusarium spp. in grain with regression-based learning algorithms

01 October 2008 → 30 September 2012
Regional and community funding: IWT/VLAIO
Research disciplines
  • Natural sciences
    • Systems biology
fusarium spp
Project description

Fusarium head blight on small grain cereals is caused by a complex disease of various Fusarium spp. whose symptoms are indistinguishable. Besides phytopathological damage that causes the disease, there is a risk of contamination of the infected grain with mycotoxins. The concern of the European Union for the presence of Fusarium mycotoxins in cereals was materialized in 2006 by the introduction of a standard for the main mycotoxin viz. Deoxynivalenol (DON). Arfusarium in Flanders is a difficult problem to manage. Due to the large variability in appearance, the prediction of a aarfusariumaantasting is not easy and not be realized with simple weighting based or logistic regression methods. Indeed, the damage is the result of many variables that may be mutually correlated. Yet it is necessary from the viewpoint of the farmer to estimate whether the cultivation of corn increased risk of Fusarium and / or seen DON contaminated corn exists possibly no longer marketable. The main objective of this research proposal is to implement learning algorithms for the prediction of Fusarium and DON in grains in the advisory system EPIPRE. EPIPRE is an advisory system that wheat growers are guided individually in Belgium and France to obtain an optimal financial return. Based on the predicted risk EPIPRE should then be able to make the project results input for preventive management advice both at the start of the cultivation in the course of cultivation. The prediction model must also provide a more adequate and targeted sampling in the detection of mycotoxins, both at the level of the grower and the trade and this in the context of IKKB programa's.