不同来源玉米对大恒肉鸡代谢能的测定及近红外预测模型的构建

Determination and NIRS Prediction of Corn Metabolizable Energy Value for Daheng Broiler

本试验旨在评定30种不同来源玉米对大恒肉鸡的代谢能值,并利用近红外光谱技术建立近红外预测模型。试验采用单因素完全随机设计,每个玉米样品为1个处理,每个处理8个重复,每个重复1只鸡,采用17周龄大恒肉公鸡。采用排空强饲法评定代谢能;结合近红外光谱,建立大恒肉鸡的玉米代谢能预测模型。结果表明:1)30种玉米的代谢能(DM)为86.76%±0.56%;粗蛋白(CP)为7.99%±0.43%;总能(GE)为16.239±0.123 MJ/kg;中性洗涤纤维(NDF)为11.27%±1.89%;酸性洗涤纤维(ADF)为2.80%±0.39%;粗纤维(CF)为1.98%±0.23%。2)30种玉米对大恒肉鸡的表观代谢能(AME)为14.624±0.469 MJ/kg;氮校正表观代谢能(AMEn)为14.646±0.462 MJ/kg;真代谢能(TME)为16.062±0.488 MJ/kg;氮校正真代谢能(TMEn)为16.083±0.481 MJ/kg。3)用近红外技术建立的大恒肉鸡AME、AMEn、TME、TMEn校正模型决定系数(R2cal)、预测标准差(RMSEE)及相对标准差(RSD)分别为0.99、0.023、0.16;0.99、0.011、0.07;0.99、0.014、0.08;0.99、0.013、0.08;内部交叉验证决定系数(R2cv)及内部交叉验证预测标准差(RMSECV)、相对标准差(RSD)分别为0.93、0.057、0.39;0.95、0.048、0.33;0.94、0.056、0.35;0.95、0.055、0.34。结果表明不同来源玉米对大恒肉鸡的AME、AMEn、TME、TMEn均存在差异。2)构建的近红外预测模型可以较好地预测大恒肉鸡的玉米代谢能值。

The study was conducted to compare the ME of 30 kinds of corns from different areas in China for Daheng broiler.And the ME predictive models of corn for Daheng broiler was setted up by using the near infrared spectroscopy technology.A completely randomized design was used and each sample of the 30 corns was regard as a treatment,with8 replicates per treatment and 1 broiler per replicate.Daheng broilers were 18-wk-old of similar weight（3.858 ±0.286kg）.Sibbald’s empty-force-feeding method was used to investigate the metabolizable energy.Meanwhile the spectra were obtained for each sample by using NIRS,and the corn ME prediction NIRS model was established on the basis of determining the ME of 30 corns from different sources.Then,the calibration and validation were performed.The result showed as follows：1） The values of DM,CP,GE,NDF,ADF,CF of 30 kinds of corns were 86.76 %± 0.56%,7.99%±0.43%,16.239 ±0.123MJ/kg,11.27% ±1.89%,2.80% ±0.39%,1.98% ±0.23%,repectively.2） The values of AME,AMEn,TME,TMEn of 30 kinds of corns for Daheng broiler were 1 4.624 ± 0.469 MJ /kg,14.646 ±0.462 MJ/kg,16.062 ±0.488 MJ/kg,16.083 ±0.481 MJ/kg,respectively.3） For the prediction modles of AME,AMEn,TME,TMEn established with near infrared reflectance spectroscopy for Daheng broiler,the coefficients of determination of calibration（R2cal）,the standard deviation of calibration（RMSEE） and relative standard deviation（RSD）of AME,AMEn,TME,TMEn were 0.99,0.023,0.16;0.99,0.011,0.07;0.99,0.014,0.08;0.99,0.013,0.08,respectively.While the coeffecients of determination of internal cross validation（R2cv）,the standard deviation of internal cross validation（RMSECV） and relative standard deviation（RSD） were 0.93,0.057,0.39;0.95,0.048,0.33;0.94,0.056,0.35;0.95,0.055,0.34,respectively.These results indicate that：1） The AME,AMEn,TME,and TMEn of corns were varied for Daheng broiler.2） It is feasible to predict AME,AMEn,TME,TMEn of corns with NIRS model for Daheng broiler.