Prediction of beef tenderness by near infrared spectroscopy

Meat tenderness has been commonly evaluated by Warner-Bratzler Shear Force (WBSF) method. This device provides a measure of the force required to shear a piece of meat. However, as this analyze is time consuming, the near infrared spectroscopy (NIRS) may be one of the most promising techniques for meat quality evaluation, as it is easy and fast to perform. The aim of this study was to evaluate the potential of NIRS based model to predict meat WBSF of Nellore cattle. Meat samples of 644 Nellore cattle, slaughtered in commercial slaughterhouse were collected between the 12th and 13th ribs of the left half-carcasses (tenderloin) from Longissimus dorsi muscle. Six spectra per sample were scanned (5.4 ⁺1.4?C) with wavelength ranging from 4,000 to 10,000 cm^-1, with accumulations of 64 scans, spectral resolution of 16 cm^-1 and data interval of 2 cm^-1, using a FT-IR Spectrum 100N (PerkinElmer, Shelton, United States). Immediately after scanning, the meat samples were analyzed for WBSF, following the standardized procedure proposed by Wheeler et al. (1995). The mean spectrum was used for each sample. In pre-processing technique, wavelength ranging from 4,132 to 9,112 cm^-1 was selected and mean centering was used. Calibration models were obtained using partial least squares regression (PLSR) with full cross validation, the same samples were used both for model estimation and testing, leaving out only one sample at a time. The models performance was assessed by the coefficient of determination of cross-validation (R²) and root mean square error of cross-validation (RMSEC). The mean and standard deviations for WBSF were 6.15 and 1.83 kg, respectively. The WBSF prediction model showed a R² of 0.40 and RMSEC of 1.42 kg. These results were equal or better than literature results for beef tenderness in different beef cattle breeds. Therefore, NIRS can be used to predict Nellore beef tenderness and can be a tool in animal breeding programs.