肉骨粉中牛羊源成分含量的近红外漫反射光谱分析

ANALYSIS OF CATTLE AND SHEEP CONTENT IN PIG OR POULTRY MEAT AND BONE MEAL BY NEAR INFRARED REFLECTANCE SPECTROSCOPY

为了切断疯牛病的传播途径保证饲料安全,开展非反刍动物肉骨粉中反刍动物成分含量的检测方法研究十分必要.本文探讨了利用近红外漫反射光谱(NIRS)分析技术快速检测非反刍动物(猪鸡)肉骨粉中反刍动物(牛羊)成分含量的可行性.收集了不同地区不同种类的肉骨粉样本,在猪和鸡肉骨粉中分别掺入0～48%的牛或羊肉骨粉,制备了200个样本.针对含牛羊源的猪肉骨粉、含牛羊源的鸡肉骨粉和含牛羊源的猪鸡肉骨粉三类样本,以掺入的牛羊源成分的质量比例作为真值,分别采用改进的偏最小二乘回归(MPLS)、偏最小二乘回归(PLS)和主成分回归(PCR)三种不同建模方法建立了NIRS定标模型.用独立的验证集对模型进行了验证,最优模型验证集的决定系数(R2)均大于0.90,相对分析误差(RPD)均大于3.0.结果表明,利用NIRS分析技术可以快速检测出猪鸡肉骨粉中牛羊源成分的含量.

In order to protect the feed chain from contamination by bovine spongiform encephalopathy （ BSE）, it is necessary to detect the content of cattle and sheep meat and bone meal （MBM） in poultry or pig MBM. The objective of this study was to investigate the feasibility to determine the content of cattle and sheep MBM in poultry or pig MBM by using near infrared reflectance spectroscopy （NIRS）. The MBM specimens were collected from different areas of China. Two hundred samples were prepared by poultry or pig MBM deliberately adulterated with cattle or sheep MBM over the weight range 0～48%. Three regression methods were compared （modified partial least squares, partial least squares and principal component regression）. Three optimum models were developed from calibration sets of poultry, pig and both, respectively. The coefficients of determination （ R2 ） and the relative prediction deviation （RPD） of independent validation sets were greater than 0.90, 3.0, respectively. It is concluded that NIRS can be used as a rapid method to detect the content of cattle and sheep MBM in poultry or pig MBM.