Prediction of beef color using time-domain nuclear magnetic resonance (TD-NMR) relaxometry data and multivariate analyses

Abstract

Time-domain nuclear magnetic resonance and chemometrics were used to predict color parameters, such as lightness (L*), redness (a*), and yellowness (b*) of beef (Longissimus dorsi muscle) samples. Analyzing the relaxation decays with multivariate models performed with partial least-squares regression, color quality parameters were predicted. The partial least-squares models showed low errors independent of the sample size, indicating the potentiality of the method. Minced procedure and weighing were not necessary to improve the predictive performance of the models. The reduction of transverse relaxation time (T2) measured by Carr–Purcell–Meiboom–Gill pulse sequence in darker beef in comparison with lighter ones can be explained by the lower relaxivity Fe2+ present in deoxymyoglobin and oxymyoglobin (red beef) to the higher relaxivity of Fe3+ present in metmyoglobin (brown beef). These results point that time-domain nuclear magnetic resonance spectroscopy can become a useful tool for quality assessment of beef cattle on bulk of the sample and through-packages, because this technique is also widely applied to measure sensorial parameters, such as flavor, juiciness and tenderness, and physicochemical parameters, cooking loss, fat and moisture content, and instrumental tenderness using Warner Bratzler shear force.