Abstract
Protein determination is of importance to characterise insect composition and quality. While different factors (kp) to convert nitrogen to protein have been proposed to report the concentration of crude protein (CP) in insect applications, the N × 6.25 is still utilised by many commercial laboratories. Different data-dependent conversion factors have been proposed to report the concentration of CP in different insect species, including black soldier fly larvae (BSFL). The objective of this paper was to evaluate the effect of using different nitrogen to protein conversion factors (kp) to predict CP in BSFL from different commercial sources using near-infrared spectroscopy. The coefficient of determination in cross-validation (R(2) (CV)) and the standard error in cross-validation (SECV) for the prediction of CP% in the BSFL were 0.75 (SECV, 4.51%), 0.75 (SECV, 4.03%), and 0.75 (SECV, 3.43%), using the kp_6.25, kp_5.65, and kp_4.76, respectively. The study showed that the different kp used can affect the cross-validation statistics (SECV) for the prediction of CP in BSFL using NIR spectroscopy. Additionally, the lower accuracies obtained for the prediction of CP are not only associated with the kp used to calculate the CP but also with the amount of chitin in the BSFL. Understanding the variables, such as reference data, that influence the calibration results using NIR spectroscopy is of importance to better provide consistent QC methods for the industry. The limitations of this study are the few numbers of samples used to develop the calibration models, although different waste streams and larvae stages were evaluated.