青脚麻肉鸡对不同来源玉米的代谢能值及近红外预测模型的构建

摘要

The study was conducted to evaluate the metabolizable energy values of 30 kinds of corns from dif⁃ferent areas in China for Qingjiaoma broilers and establish the metabolizable energy predictive models of corn by using the near infrared spectroscopy ( NIRS) technology, in order to accumulate the basic data for building feed nutritional value database and fast predicting metabolizable energy of corn for Qingjiaoma broilers. A sin⁃gle⁃factor completely random design was used in this trial, and a total of 48 30⁃week⁃old Qingjiaoma broilers with similar body weight were randomly assigned into several groups for 9 patches of metabolism experiments with 3 or 4 corn samples each, 8 replicates in each sample and 1 broiler in each replicate. An endogenous group was prepared in every metabolism experiment, and 10⁃day recovery between each batch was set up. Sib⁃bald’ s empty⁃force⁃feeding method was used to investigate the metabolizable energy, including feed withdraw⁃al for 48 h before feeding test samples and then 2% of body weight of corn for force feeding, with 48 h of ex⁃creta collection period;whereas the group for endogenous collection was continued to fast and collect the excre⁃ta for 48 h. The result showed as follows:1)the dry matter content in 30 corn samples was (86.75±0.55)%(85.55% to 87.79%), and based on the dry matter, crude protein content, gross energy, contents of neutral detergent fiber, acid detergent fiber, crude fiber were ( 9. 21 ± 0. 52 )% ( 8. 27% to 10. 58%) , ( 18. 716 ± 0.106) MJ/kg (18.429 to 18.951 MJ/kg), (13.00±2.21)% (10.00% to 18.52%), (3.23±0.46)%(2.37% to 4.36%) and (2.28±0.28)% (1.89% to 2.76%), respectively. 2) Based on the dry matter, ap⁃parent metabolizable energy ( AME) , nitrogen corrected apparent metabolizable energy ( AMEn) , true metab⁃olizable energy ( TME) and nitrogen corrected true metabolizable energy ( TMEn) of 30 corns for Qingjiaoma broilers were (14.627±0.655) MJ/kg (11.727 to 16.225 MJ/kg), (14.672±0.641) MJ/kg (11.793 to 16.248 MJ/kg) , ( 16. 248 ± 0. 619 ) MJ/kg ( 13. 333 to 17. 727 MJ/kg ) and ( 16. 293 ± 0. 605 ) MJ/kg ( 13.398 to 17.750 MJ/kg) , respectively. 3) The predictive models of AME, AMEn, TME and TMEn were established with NIRS for Qingjiaoma broilers, and the determination coefficients of calibration ( R2cal ) , the standard deviation of calibration ( RMSEE) and relative standard deviation ( RSD) of AME, AMEn, TME, TMEn were 0.99, 0.035, 0.24;0.99, 0.029, 0.20;0.99, 0.031, 0.19;and 0.99, 0.030, 0.18;respective⁃ly. Meanwhile, the determination coefficients of cross validation (R2cv), the standard deviation of cross valida⁃tion ( RMSECV) and RSD were 0.92, 0.117, 0.80;0.93, 0.106, 0.73;0.90, 0.113, 0.70;0.91, 0.108, 0.66; respectively. These results indicate that:1) The AME, AMEn, TME and TMEn of different corns are varied for Qingjiaoma broilers. 2 ) It is feasible to predict corn metabolizable energy with NIRS model for Qingjiaoma broilers.%本试验旨在评定青脚麻肉鸡对30个不同来源玉米的代谢能值，并利用近红外光谱技术构建代谢能值的预测模型，为构建青脚麻肉鸡饲料营养价值数据库和玉米代谢能值的快速预测积累基础数据。试验采用单因素完全随机设计，选用48只体重相近的30周龄青脚麻肉公鸡，按照组间体重无差异原则随机分组，做9批次代谢试验，每批做3或4个玉米样，每样设8个重复，每个重复1只鸡；每批做1个内源组，每批之间设10 d恢复期。试验采用排空强饲法测定代谢能值，试鸡饥饿48 h，然后按体重2％强饲待测饲料，收集排泄物48 h；内源组鸡饥饿48 h，再继续饥饿收集排泄物48 h。结果显示：1）30个玉米样品的干物质含量为（86．75±0．55）％（85．55％～87．79％），以干物质为基础，粗蛋白质含量为（9．21±0．52）％（8．27％～10．58％），总能为（18．716±0．106） MJ／kg（18．429～18．951 MJ／kg），中性洗涤纤维含量为（13．00±2．21）％（10．00％～18．52％），酸性洗涤纤维含量为（3．23±0．46）％（2．37％～4．36％），粗纤维含量为（2．28±0．28）％（1．89％～2．76％）。2）以干物质为基础，青脚麻肉鸡对30种玉米的表观代谢能（AME）为（14．627±0．655） MJ／kg（11．727～16．225 MJ／kg），氮校正表观代谢能（AMEn）为（14．672±0．641） MJ／kg（11．793～16．248 MJ／kg），真代谢能（TME）为（16．248±0．619） MJ／kg （13．333～17．727 MJ／kg），氮校正真代谢能（TMEn）为（16．293±0．605） MJ／kg（13．398～17．750 MJ／kg）。3）用近红外光谱技术建立的青脚麻肉鸡AME、AMEn、TME、TMEn校正决定系数（R2cal）、校正标准差（RMSEE）及相对标准差（RSD）分别为0．99、0．035、0．24，0．99、0．029、0．20，0．99、0．031、0．19，0．99、0．030、0．18；交叉验证决定细数（R2cv）、交叉验证标准差（RMSECV）及RSD分别为0．92、0．117、0．80，0．93、0．106、0．73，0．90、0．113、0．70，0．91、0．108、0．66。结果表明：1）青脚麻肉鸡对不同来源玉米的AME、AMEn、TME和TMEn存在差异；2）近红外模型可以较好地预测青脚麻肉鸡的玉米代谢能值。