TRALSPEC-AI |Development and validation of a novel method for the determination of Tropane Alkaloids in Food and Food Products

Datura spp. is rich in tropane alkaloids (TAs), plant secondary compounds produced in response to environmental stressors. The entire plant is toxic and now invade crops such as maise, millet, amaranth, buckwheat, flax/linseed, sunflowers, sorghum, and soybeans. Recently, many people in Uganda suffered from severe food poisoning illness and fatality due to the consumption of Super Cereal contaminated by TAs. The gold standard methods for detecting and quantifying TAs in foods are time-consuming, laborious, expensive. They require a high level of expertise-- implying that stakeholders, such as the crop producers, cannot use these in quality control of their commodities. Our proposal aims to answer the question: can TA contamination that can poison the consumers be detected and thus prevented using vibrational spectroscopy coupled with data analytics to give accurate and real-time measurements? To answer this question, we aim to develop and validate a novel approach to gold standard methods in quantifying TAs in foods, therefore alerting for toxicity. The methods will be based on validated vibrational spectroscopic (infrared) techniques. Foods with a wide range of TA concentrations will be obtained from the European Union Reference Laboratories. They will be scanned using a benchtop and portable IR instrument. They will use Artificial intelligence (Machine Learning) to analyse the large volume of spectral data, resulting in predictive modelling that will detect concentrations of TAs in scanned foods. The model will be imported into the benchtop and portable IR instruments, thus in a way, automating the TA analysis to give real-time measurements. This is user-friendly and can be used at any stage along the food supply chain. We will compare the results obtained from this novel approach with the gold standard methods (Gas & Liquid chromatography coupled with high-resolution mass spectrometry (MS) or MS/MS).