Embrapa Instrumentação, Rua Quinze de Novembro 1452, São Carlos, SP 13561-206, Brazil.
Talanta. 2013 Apr 15;108:88-91. doi: 10.1016/j.talanta.2013.02.070. Epub 2013 Mar 7.
The noteworthy of this study is to predict seven quality parameters for beef samples using time-domain nuclear magnetic resonance (TD-NMR) relaxometry data and multivariate models. Samples from 61 Bonsmara heifers were separated into five groups based on genetic (breeding composition) and feed system (grain and grass feed). Seven sample parameters were analyzed by reference methods; among them, three sensorial parameters, flavor, juiciness and tenderness and four physicochemical parameters, cooking loss, fat and moisture content and instrumental tenderness using Warner Bratzler shear force (WBSF). The raw beef samples of the same animals were analyzed by TD-NMR relaxometry using Carr-Purcell-Meiboom-Gill (CPMG) and Continuous Wave-Free Precession (CWFP) sequences. Regression models computed by partial least squares (PLS) chemometric technique using CPMG and CWFP data and the results of the classical analysis were constructed. The results allowed for the prediction of aforementioned seven properties. The predictive ability of the method was evaluated using the root mean square error (RMSE) for the calibration (RMSEC) and validation (RMSEP) data sets. The reference and predicted values showed no significant differences at a 95% confidence level.
本研究的特点是使用时域核磁共振（TD-NMR）弛豫度数据和多元模型来预测牛肉样本的七个质量参数。将 61 头邦斯玛拉小母牛的样本根据遗传（繁殖组成）和饲料系统（谷物和草料饲料）分为五组。采用参考方法对七个样品参数进行分析；其中，三个感官参数，风味、多汁性和嫩度，以及四个理化参数，烹饪损失、脂肪和水分含量以及使用 Warner-Bratzler 剪切力（WBSF）的仪器嫩度。对同一动物的生牛肉样本进行 Carr-Purcell-Meiboom-Gill（CPMG）和连续波自由进动（CWFP）序列的 TD-NMR 弛豫度分析。使用偏最小二乘（PLS）化学计量技术，基于 CPMG 和 CWFP 数据以及经典分析的结果构建回归模型。结果允许预测上述七个性质。使用校准（RMSEC）和验证（RMSEP）数据集的均方根误差（RMSE）评估方法的预测能力。在 95%置信水平下，参考值和预测值没有显著差异。
