Prediction of chevon quality through near infrared spectroscopy and multivariate analyses

The aim of this study was to test the ability of near-infrared (NIR) reflectance spectroscopy to predict dry matter, crude protein, ether extract, ash, moisture, cooking loss, and drip loss of chevon . In total, 114 samples were collected from 38 young (two teeth aged) castrated goat carcasses from a local market in Mymensingh district of Bangladesh. For conducting the studyExperimental longissimus dorsi (LD) muscle were sampled from 9th to 13th ribs in the early morning hours. A total of 342 NIRs spectra were collected using the DLP NIRscan Nano Software and average spectrum was 114. Partial least square regression analysis for the calibration and validation models were developed using the Unscrambler X software. Prediction models were satisfactory for dry matter (R2 = 0.75), crude protein (R2 = 0.82), moisture (R2 = 0.75), and drip loss (R2 = 0.83). The most promising model found for ash (R2 = 0.85), and Root Mean Square Errors (RMSE) also very low (0.15). Lowest R2 was found for cooking loss at 0.57. Based on these results, the NIR spectroscopy and multivariate analysis method were reasonably efficient for the rapid assessment of physicochemical traits of ash, drip loss, crude protein, moisture, and dry matter content of chevon.