Starch and moisture interfere with fungal detection

Aflatoxins are highly carcinogenic toxins that can can cause liver damage in humans, especially children. They are produced by certain Aspergillus species of fungi, particularly A. flavus and A. parasiticus, which are prone to colonising food grains such as corn, peanuts and rice during storage, especially in humid conditions. In order to protect consumers, regulatory authorities such as the European Commission set thresholds for aflatoxin contamination in food grains, requiring agricultural producers to be on a constant look-out for fungal infections.

Detecting infection by Aspergillus species in stored grain usually involves trying to culture the fungi in the laboratory or detect their DNA using the polymerase chain reaction, both of which are complex, time-consuming activities. Recently, however, several groups have shown that NIR spectroscopy offers a quicker and simpler method for detecting Aspergillus infection and aflatoxin contamination in various different grains, including maize and barley. Inspired by this work, a team of scientists from Thailand has now tried using NIR spectroscopy to detect Aspergillus infection in rice, but with mixed results.

As they report in Food Control, the scientists analysed over 100 rice samples, all of which were infected with Aspergillus to a greater or lesser extent, with NIR spectroscopy. Their aim was to try to determine levels of both total fungal infection and infection with aflatoxin-producing species.

They found that the NIR signals at certain wavelengths did seem to correlate with levels of total fungal infection and infection with aflatoxin-producing species, but this was complicated by the fact that the signals also correlated with the moisture and starch content of the rice grains. You would expect a greater degree of infection in rice grains with higher levels of starch and moisture, so it may be that the NIR signals were mainly responding to variations in the moisture and starch levels rather than to the extent of infection.

Indeed, when the scientists developed models for predicting fungal infection from the NIR data, they weren't particularly accurate. Nevertheless, the scientists think this approach still has merit, but that future studies will need to take into account the influence of the starch and moisture content on the NIR signal.

Blog tags: