Rapid prediction of beef quality remains a challenge for meat processors. This study evaluated the potential of Raman spectroscopy followed by chemometrics for prediction of Warner-Bratzler shear force (WBSF), intramuscular fat (IMF), ultimate pH, drip-loss and cook-loss. PLS regression models were developed based on spectra recorded on frozen-thawed day 2 longissimus thoracis et lumborum muscle and validated using test sets randomly selected 3 times. With the exception of ultimate pH, models presented notable performance in calibration (R² ranging from 0.5 to 0.9; low RMSEC) and, despite variability in the results, promising predictive ability: WBSF (RMSEP ranging from 4.6 to 9 N), IMF (RMSEP ranging from 0.9 to 1.1%), drip-loss (RMSEP ranging from 1 to 1.3%) and cook-loss (RMSEP ranging from 1.5 to 2.9%). Furthermore, the loading values indicated that the physicochemical variation of the meat influenced the models. Overall, results indicated that Raman spectroscopy is a promising technique for routine quality assessments of IMF and drip-loss, which, with further development and improvement of its accuracy could become a reliable tool for the beef industry.

快速预测牛肉质量对肉类加工者仍然是一个挑战。这项研究评估了拉曼光谱法和随后的化学计量学方法预测Warner-Bratzler剪切力（WBSF），肌内脂肪（IMF），最终pH，滴漏和蒸煮损失的潜力。基于在第2天冻融后的胸大肌和腰部肌肉上记录的光谱开发PLS回归模型，并使用随机选择3次的测试集进行验证。除极限pH值外，模型在校准（R ²0.5至0.9; 较低的RMSEC），尽管结果存在差异，但具有良好的预测能力：WBSF（RMSEP范围从4.6到9 N），IMF（RMSEP范围从0.9到1.1％），滴漏（RMSEP范围从1到1.3％）和蒸煮损失（RMSEP范围为1.5至2.9％）。此外，负荷值表明肉的理化变化影响了模型。总体而言，结果表明，拉曼光谱法是用于IMF和滴漏常规质量评估的有前途的技术，随着进一步的发展和准确性的提高，拉曼光谱法可以成为牛肉行业的可靠工具。