Development of Representative Sampling Plans for Mycotoxins in Foods Using Distribution Modelling

Initially, a review of the scientific literature will be carried out to identify sources of data that can be used to create the model (including the primary data associated with previous studies, sampling Directives and protocols).
<p>
Using the information gathered, a computer program will be written to produce a pilot model of distribution and sampling for one mycotoxin and one commodity. The model will generate probability distributions for the concentrations of toxins throughout a grain store, using at least two variables to characterise the distribution of the contamination in the model, including 1) the 'patch' size and 2) the 'patch concentration'.
<p>
Once the pilot model has been established, the effects of processing on mycotoxin distribution will be incorporated. Simple transformation of the original parameters will show the effect that mixing or separation of the commodity can have on the distribution of the contamination, such operations can cause concentration or dilution of the mycotoxin in the system.
<p>
Finally, the mechanism to test sampling strategies will be added to the model, established sampling protocols will be assess using the model and the effectiveness of new sampling schemes (e.g. the effect of varying the number of samples, their position and incremental sample size) will be tested.