Using near infrared spectroscopy to predict the physical traits of Bos grunniens meat

The purpose of the current study was to evaluate the use of near infrared spectroscopy to predict the physical traits of yak (Bos grunniens) meat. Near infrared spectra data of M. longissimus thoracis from 162 yaks were collected, as well as physical traits data. Prediction models of physical traits using near infrared spectroscopy were established through partial least squares regression (PLSR) combined with mathematical pretreatments such as orthogonal signal corrections (OSC) and detrending corrections (DT). The coefficients of determination of calibration (RC2) of prediction models were higher than 0.6 except WBSF. The ratio performance deviation for a* and b* values and their derived values SI and HA exceeded 2.0, which are suitable for screening. Through modeling separately for steers and heifers, we found that prediction ability for cooking loss would be increased but for b* would be decreased. Conclusively, near infrared spectroscopy is useful for predicting physical traits of yak meat and could be a potential tool for monitoring beef quality of the yak.