The host resistance strategy for eliminating aflatoxins from corn has been advanced by the discovery of natural resistance traits such as proteins. This progress was aided by the development of a rapid laboratory-based kernel screening assay (KSA) used to separate resistant from susceptible seed, and for investigating kernel resistance.A. flavus GUS transformants have also been used, in conjunction with the KSA, to assess the amount of fungal growth in kernels and compare it with aflatoxin accumulation. Several proteins associated with resistance (RAPs) have been identified using 1 D PAGE. However, proteomics is now being used to further the discovery of RAPs. This methodology has led to the identification of stress-related RAPs as well as other antifungals. Characterization studies being conducted, including RNAi gene silencing experiments, may confirm roles for RAPs in host resistance.

In this paper we start from a critical survey of building expert systems from the perspective of knowledge acquisition. We propose a novel approach in which knowledge is acquired through mental models, and the expert systems are built by integrating these mental models. The architecture of this kind of system is described, some important features of this approach are also discussed.