Near-infrared (1100-2500 nm) reflectance spectra were measured on 128 beef ribeye samples using a scanning monochromometer and calibrated against standard laboratory analysis to develop a rapid method for measuring intramuscular fat of beef. A pilot run was conducted on 38 frozen ribeye samples, ranging in fat content from 94 to 325g/kg dry weight, to evaluate sample preparation methods (frozen slices, thawed slices, coarse mince. After scanning of subsamples, fat content of the sample was then measured using Soxhlet extraction. Standard error of calibration (SEC) was 9g/kg dry weight for spectra collected from minced samples compared to 25g/kg for frozen slices and 15g/kg for thawed slices. A second set of 90 frozen ribeye samples (fat content range 56-305 g/kg dry weight) was used to test and improve this calibration, using 3 subsamples of minced beef per sample. The ability of the initial calibration equation to predict fat content of these samples was poor. A more robust calibration was then attempted by pooling the data from the two sets. The SEC of this extended set was 16 g/kg dry weight (10% of the sample set mean) and the multiple R2 of calibration was 0.92. The standard error of prediction for new samples was estimated from internal cross-validation runs as 19 g/kg. Prediction of fat content of an independent validation set of 24 samples had a standard error of 20 g/kg dry weight. Analysis of the calibration results suggested that a major source of error in the calibration was hererogeneity between subsamples, with the standard deviation of prediction of different subsamples from the same ribeye being larger than the standard error of cross validation. These results indicate that Near-Infrared reflectance spectroscopy may provide a rapid objective method for estimating intramuscular fat of beef.

TR, Mackle, CR Parr, and AM Bryant

Proceedings of the New Zealand Society of Animal Production, Volume 55, , 82-84, 1995
Download Full PDF BibTEX Citation Endnote Citation Search the Proceedings

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.