Portable vibrational spectroscopic methods can discriminate between grass-fed and grain-fed beef

Abstract

The present study analysed the ability for portable near infrared reflectance (NIR) and Raman spectroscopy sensors to differentiate between grass-fed and grain-fed beef. Scans were made on lean and fat surfaces of 108 beef steak samples labelled as grass-fed (n = 54) and grain-fed (n = 54), with partial least squares discriminant analysis (PLS-DA) and linear discriminant analysis (LDA) used to develop discrimination models which were tested on independent datasets. Furthermore, PLS-DA was used to predict visual marbling score and days on feed (DOF). The NIR spectra accurately discriminated between grass- and grain-fed beef on both fat (91.7%, n = 92) and lean (88.5%, n = 96), as did Raman (fat 95.2%, n = 82; lean 69.6%, n = 68). Fat scanning using NIR spectroscopy moderately predicted DOF (r²_val = 0.53), though Raman and NIR spectroscopy lean prediction models for DOF and marbling were less precise (r²_val < 0.50). It can be concluded that portable NIR and Raman spectrometers can be used successfully to differentiate grass-fed from grain-fed beef and therefore aid retail and consumer confidence.