Effects of grain sampling procedures on fusarium mycotoxin assays in wheat grains.

Fusarium mycotoxins are increasingly studied agronomically, chemically, and pathologically in the context of food safety, as a means of preventing new major health crises. Reliable mycotoxin techniques and sampling procedures are required for assessment of the effects of different sources of variation on grain mycotoxin content in agronomic experiments. Analyses were performed with the aim of formulating guidelines for grain sampling to increase the reliability of grain mycotoxin measurement in agronomic experiments. Two toxins in wheat samples, deoxynivalenol and nivalenol, were targeted. With a nested linear mixed model, it was estimated that the uncertainty of nivalenol determination was low (+/-15 microg/kg), whereas that for deoxynivalenol determination was higher (+/-38 microg/kg). It was also found that grinding of the grain decreased the variability of the results. Moreover, despite the heterogeneity in grain mycotoxin content across a given field, it was shown that heads can be harvested manually for agronomic experiments provided that sampling is representative (evenly distributed over the entire plot area). Finally, delaying the assay until after harvest was found to affect the results obtained and should therefore be avoided.