摘要
本试验旨在基于康奈尔净碳水化合物-蛋白质体系(CNCPS)建立北京市全株玉米原料营养成分数据库,并利用近红外光谱(NIRS)方法建立其营养价值预测模型。试验采集北京市18个牧场89份全株玉米原料样品,测定其营养成分,利用CNCPS 6.5计算各样品碳水化合物(CHO)和蛋白质组成。定标集和验证集根据4∶1的配比关系,分别选用71份和18份全株玉米原料样品作为定标集和验证集评价NIRS模型。结果显示:1) NIRS分析技术对全株玉米原料常规营养成分、CNCPS中蛋白质组成和CHO组成均具有较好的预测能力,且精确度较高。2)干物质(DM)、粗灰分(Ash)、粗蛋白质(CP)、粗脂肪(EE)、中性洗涤纤维(NDF)、酸性洗涤纤维(ADF)、酸性洗涤木质素(ADL)、淀粉(Starch)、中性洗涤不溶蛋白质(NDIP)、酸性洗涤不溶蛋白质(ADIP)、可溶性蛋白质(SP)、CHO、非纤维性碳水化合物(NFC)、可溶性纤维(CB2)、可消化纤维(CB3)、不消化纤维(CC)、可溶性真蛋白质(PA2)、难溶性真蛋白质(PB1)、纤维结合蛋白质(PB2)和非降解蛋白质(PC)的定标决定系数(1-VR)均>0.80,验证决定系数(RSQv)均≥0.84,这些模型均可用于日常快速检测分析。DM、Ash、EE、NDF、ADF、ADL、Starch、NDIP、CHO、NFC、CB2、CB3、PC和PB1的NIRS模型参数均采用二阶导数处理,CP、SP、ADIP、CC、PA2和PB2的NIRS模型参数均采用标准正态变量+二阶导数处理。综上所述,本研究提供了全株玉米原料的基础化学分析数据,并通过NIRS分析技术建立了主要营养成分的快速预测模型,有利于养殖场青贮前对全株玉米原料质量的快速评估。
This study aimed to establish a database of nutrient components of whole-plant corn raw material in Beijing based on the Cornell net carbohydrate and protein system(CNCPS),and found nutrient value prediction models using the method of near infrared reflectance spectroscopy(NIRS).A total of 89 whole-plant corn raw material samples were collected from 18 dairy farms in Beijing,and the nutrient components were determined,then the carbohydrate(CHO)and protein components were calculated by CNCPS 6.5.The calibration set and verification set were based on a 4∶1 ratio,and NIRS models were evaluated using 71 and 18 samples of whole-plant corn raw material as calibration and validation database,respectively.The results showed as follows:1)the conventional nutrient components of whole-plant corn raw material,protein composition and CHO composition in CNCPS system could be quite accurately estimated by NIRS technology.2)The cross validation determinant coefficients(1-VR)were>0.8,and the verification decision coefficients(RSQv)were≥0.84 for the model parameters of dry matter(DM),crude ash(Ash),crude protein(CP),ether extract(EE),neutral detergent fiber(NDF),acid detergent fiber(ADF),acid detergent lignin(ADL),starch(Starch),neutral detergent insoluble protein(NDIP),acid detergent insoluble protein(ADIP),soluble protein(SP),CHO,non fiber carbohydrates(NFC),soluble fiber(CB2),digestible fiber(CB3),indigestible fiber(CC),soluble true protein(PA2),insoluble true protein(PB1),fiber conjugated protein(PB2)and undegradable protein(PC),which suggested that these models could be used for rapidly actual analysis.Model parameters of DM,Ash,EE,NDF,ADF,ADL,Starch,NDIP,CHO,NFC,CB2,CB3,PC and PB1 were processed by second derivative,and CP,SP,ADIP,CC,PA2 and PB2 were processed by standard normal variate+second derivative.In conclusion,the study provide chemical analysis data of hole-plant corn raw material,and establish models for prediction of main nutrient components by NIRS technology.It is beneficial to the evaluation of the quality
作者
刘娜
屠焰
刁其玉
郭江鹏
齐志国
司丙文
王俊
吴万成
陈国顺
LIU Na;TU Yan;DIAO Qiyu;GUO Jiangpeng;QI Zhiguo;SI Bingwen;WANG Jun;WU Wancheng;CHEN Guoshun(Gansu Agricultural University,Lanzhou 730070,China;Key Laboratory of Feed Biotechnology of the Ministry of Agriculture,Feed Research Institute of Chinese Academy of Agricultural Sciences,Beijing 100081,China;Beijing Municipal Animal Husbandry Station,Beijing 100107,China)
出处
《动物营养学报》
CAS
CSCD
北大核心
2019年第5期2287-2295,共9页
CHINESE JOURNAL OF ANIMAL NUTRITION
基金
奶牛产业技术体系北京市创新团队(BAIC06-2018)
中国农业科学院科技创新工程协同创新任务"奶牛绿色养殖技术集成创新"(CAAS-XTCX2016011-01)
