Authenticating common Australian beef production systems using Raman spectroscopy

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

“…However, its' ability to discriminate correctly following fat scanning was comparable to NIR spectra after data cleaning, which was required due to the high variability of Raman spectra within and between samples. The PLS-DA accuracy of Raman in the present study was comparable to recent findings scanning fat (85–96%), 22 although lean was less accurate for discrimination purposes, being 69.5% compared to 96.5%–98.5%. 13,21 Similarly, precision for fat using PLS-DA in both the present study and Logan et al 22 was 100%.…”

“…It is therefore recommended that future work should use larger datasets with staggered DOF samples across a wider range, such as 0, 70, 100, 150, 300, 450 and 600 DOF. For instance, a recent study using Raman scanning subcutaneous fat was able to classify beef ( n = 520) into different DOF groups (grass, grass + supplement, 70 and 100 DOF) at 87.1% accuracy and 86.8% precision by multi-class PLS-DA instead of PLS regression, 22 as was used in the present study. Fat proved to be a better predictor for DOF than lean for Raman, which may have occurred due to changes in the fatty acid profile with DOF.…”

“…The present study attempted to predict the length of time animals were fed a high grain diet in a feedlot through DOF which is an important aspect for many consumers when making purchase decisions as it often infers Wagyu genetics and increased fat content, marbling and desirable flavours associated with these. 8,22,60,61 The predictions of DOF and marbling were made with the objective to infer whether fat content would lead to the high accuracy of discriminating grass- and grain-fed beef. However, the predictions of DOF were poor in the present study except for scanning of fat using the Raman spectrometer, which showed moderate precision.…”

“…Fat proved to be a better predictor for DOF than lean for Raman, which may have occurred due to changes in the fatty acid profile with DOF. 8,22 Fat colour would be expected to be lighter from longer grain feeding durations due to decreased β-carotene derived from pasture feeding. 9,55 However, the spectrometers used in the present study do not measure reflectance in visible range and that may explain the low accuracy for predicting DOF.…”

“…16–20 However, no evidence of studies using NIR spectroscopy of intact beef for purposes of discriminating grass-fed and grain-fed feeding regime have been found. Recent studies have shown successful discrimination analysis of feeding regime (grass-fed and grain-fed) on intact beef brisket muscle (accuracy 97.1–98.5%) 13,21 and on subcutaneous fat (accuracy 85–96%) 22 using a handheld Raman spectrometer. This evidence presents an opportunity for future use of both NIR and Raman spectroscopy in meat processing or retail sectors for authentication.…”