Assessing chemometric models developed using Raman spectroscopy and fatty acid data for Northern and Southern Australian beef production systems

Abstract

A total of 960 beef carcases from northern and southern Australian production regions were assessed by examining the subcutaneous fat. Carcases from four different production systems within each region were assessed, by Raman spectroscopy and the fatty acid composition determined to develop models that best classified the various production systems. As a result, 12 Partial Least Square Discriminant Analysis models were developed. A two-class model based on fatty acid composition was able to correctly classify 99% of grass and grain fed animals. The best Raman spectroscopic model correctly classified 94% of grass vs grain carcases produced in the northern region. For the southern production region, the models had the following classification accuracies; southern long-term grain fed (98%), southern short-term grain fed (95%), southern grass (96%), southern grass supplemented (97%), and the southern model classified grass vs grain (97%). Raman spectroscopy is considered a useful rapid method for classification of beef carcases based upon production system.

Keywords: Beef cattle; PLS-DA; Production system; Raman spectroscopy.