Machine learning approach for assessing lipid oxidation and deterioration in nutritional value in pearl millet due to Tribolium castenum infestation

Tribolium castenum infestation causes severe deterioration of pearl millet grains and flour. Pearl millet grains stored at 15, 25 and 35 °C and artificially infested with T. castenum at different stages of infestation (1st day, egg, larvae, pupa and adult) were evaluated for the nutritional value (protein and fat content) and rancidity indicators free fatty acid (FFA), peroxide value (PV) and acid value (AV). Storage temperature and infestation by T castenum had a significant effect on the above-mentioned properties. At the adult stage of infestation, protein content, FFA value, AV, and PV increased by 1.3 times, 2.4 times, 2 times, and 5 times, respectively, at 35 °C of storage temperature. On the other hand, fat content decreased by 4 factors compared to control samples at the same storage temperature. Protein, FFA, PV and AV had a strong positive correlation, while fat content showed strong negative correlation. Principal component analysis (PCA) was used to statistically analyse the quantitative data. The PCA analysis revealed that FFA, PV, and AV are positively correlated and contribute significantly to PC1, while Fat was negatively correlated with these parameters. Protein content behaved differently, contributing strongly to PC2, suggesting its degradation followed an independent pathway.