应用近红外光谱分析技术预测小麦肉鸭代谢能的研究

Study on Estimating Metabolizable Energy of Wheat for Meat Ducks by Near Infrared Reflectance Spectroscopy

本试验旨在实测不同来源小麦的肉鸭表观代谢能(AME),并利用近红外光谱分析技术(NIRS)构建其预测模型。选用1周龄的樱桃谷肉鸭410只,随机分为41个处理,每个处理5个重复,每个重复2只肉鸭,各处理肉鸭分别饲喂玉米-豆粕型基础饲粮和40种小麦替代饲粮(含20%小麦)。用套算法计算小麦的AME,然后利用NIRS建立小麦AME的预测模型。结果表明,不同来源小麦的肉鸭AME为11.03~14.34 MJ/kg,变异系数为5.58%;小麦AME与粗纤维(CF)、中性洗涤纤维(NDF)、酸性洗涤纤维(ADF)含量呈极显著负相关(P<0.01),与粗脂肪含量(EE)呈极显著正相关(P<0.01)。小麦AME的预测模型的定标决定系数、定标标准差和交叉验证相对标准差分别为0.85、0.187 MJ/kg和1.70%;外部验证决定系数、外部验证相对标准差和外部验证相对分析误差分别为0.89、1.46%和3.23%。由此可见,不同来源小麦肉鸭AM E和化学成分含量存在差异,其AM E的变异与其化学成分相关,应用NIRS预测小麦的肉鸭AM E的结果"良好"。

This study was conducted to measured the apparent metabolizable energy(AME)of wheat for meat ducks and establish the AME predictive models of wheat by using the near infrared reflectance spectroscopy(NIRS).A total 410 one-week-old Cherry Valley meat ducks were randomly assigned to 41 treatments with 5 replicates per treatment and 2 ducks per replicate,meat ducks in the 41 treatments were fed the corn-soybean meal base diet and 40 kinds of wheat substitute diet(contained 20%wheat),respectively.The AME of wheat were measured by the substitution method,and the prediction model of wheat AME was established by NIRS.The results showed that the AME of wheat for meat ducks ranged from 11.03 to 14.34 MJ/kg,and the coeffi-cient of variation was 5.58%.The AME of wheat was significant negative correlated with crude fiber(CF),neutral detergent fiber(NDF)and acid detergent fiber(ADF)contents(P<0.01),however,the AME of wheat was significant positive correlated with ether extract(EE)content(P<0.01).The coefficients of deter-mination of calibration,standard deviation of calibration and relative standard deviation of cross-validation for prediction model of AME of wheat were 0.85,0.187 MJ/kg and 1.70%;the coefficients of determination of external validation,relative standard deviation of external validation and ratio of performance to deviation of external validation were 0.89,1.46%and 3.23%.These results indicate that the chemical composition contents and AME are varied among different sources wheat.The AME of wheat is correlated with its chemical compo-sition and it has a good predictive performance of wheat AME by NIRS prediction.