Monitoring predictive and informative indicators of the energy status of dairy cows during early lactation in the context of monthly milk recordings using mid-infrared spectroscopy

作     者：Mueller, U. Kesser, J. Koch, C. Helfrich, H. -P. Rietz, C. 

作者机构：Univ Bonn Inst Anim Sci Physiol & Hyg Unit Bonn Germany Hofgut Neumuehle Educ & Res Ctr Anim Husb Muenchweiler Germany Univ Bonn Inst Numer Simulat Bonn Germany Heidelberg Univ Mixed Methods Res Dept Educ Sci Heidelberg Germany 

出 版 物：《LIVESTOCK SCIENCE》 (畜产科学)

年 卷 期：2019年第221卷

页      面：6-14页

核心收录：

学科分类：0905[农学-畜牧学] 09[农学] 

基　　金：North Rhine-Westphalia Ministry for Climate Protection  Environment  Agriculture  Nature Conservation and Consumer Protection (MKUNLV) within the Inter-Departmental Centre of Sustainable Agriculture (USL) 

主　　题：Energy balance beta-hydroxybutyrate in serum Monthly milk recording Linear regression models 

摘      要：With the onset of lactation, voluntary feed intake does not increase as fast as the milk yield and the energy balance (EB) becomes physiologically negative. The risk of metabolic disorders increases with decreasing EB. During this time the concentration of beta-hydroxybutyrate (BHB) in serum increases. Herd managers cannot calculate the EB as the feed intake is commonly not quantified. Measuring the concentration of BHB in serum is generally not feasible on farms because blood samples must be collected and the respective analyses run. The motivation for developing the presented models was to be able to utilise the potential of the mid-infrared (MIR) spectroscopy spectral data (as part of the monthly milk recording)) for predicting the extensive energy status indicators EB and the BHB in serum in the first crucial lactation weeks (LW). In the first 10 LW, milk and blood samples were sampled from 65 primiparous (n = 20) and multiparous (n = 45) German Holstein cows. Milk samples were analysed by the laboratory of the milk recording organisation (MRO). For the reference data, blood samples were taken for BHB analysis and the EB of each cow was calculated using automatic weighing troughs for the energy intake. To predict the indicators of energy status, linear regression models that are easily applied in real-time were combined with milk spectral data, milk component data, and information on animal lactation number and stage. The linear regression model obtained better accuracy than the partial least square. In the first 4 LW the best linear regression model reached R-2 = 0.55 for EB in the multiparous cows (MC) and R-2 = 0.83 for EB in the primiparous cows (PC);the best model for the 5th to 10th LW reached R-2 = 0.57 for EB in the MC and R-2 = 0.71 for EB in the PC. To predict serum BHB in the first 4 LW the best linear regression model reached R-2 = 0.42. Sensitivity (sen) and specificity (spec) were calculated (cut-off values: BHB 1200 mu mol/L;EB 5% percentile). Cow