Reference : Potential for assessing the pregnancy status of dairy cows by mid-infrared analysis o...
Scientific congresses and symposiums : Paper published in a book
Life sciences : Animal production & animal husbandry
http://hdl.handle.net/2268/177703
Potential for assessing the pregnancy status of dairy cows by mid-infrared analysis of milk
English
Laine, Aurélie [Université de Liège - ULiège > Sciences agronomiques > Zootechnie >]
Bel Mabrouk, Hana [Université de Liège - ULiège > Sciences agronomiques > Zootechnie >]
Dale, Laura-Monica [Université de Liège - ULiège > Sciences agronomiques > Zootechnie >]
Bastin, Catherine [Université de Liège - ULiège > Sciences agronomiques > Zootechnie >]
Gengler, Nicolas mailto [Université de Liège - ULiège > Agronomie, Bio-ingénierie et Chimie (AgroBioChem) > Ingénierie des productions animales et nutrition >]
26-Aug-2014
Book of Abstracts of the 65th Annual Meeting of the European Federation of Animal Science
Wageningen Academic Publishers
Nº. 20
157
Yes
International
978-90-8686-248-1
Wageningen
The Netherlands
65th Annual Meeting of the European Association for Animal Production
25-29 August 2014
EAAP - European Federation of Animal Science
Copenhagen
Denmark
[en] Mid-infrared ; Dairy cattle ; analysis of milk
[en] In dairy cattle, in opposition to other species, performances recording schemes allow to provide advisory tools that integrate information across the whole population. Mid-infrared (MIR) analysis of milk provides a spectrum for each individual cow’s milk sample. The MIR spectrum represents the whole milk composition and can be used to assess the status of the animal (e.g. health, pregnancy, feeding). The main objective of the European project OptiMIR (INTERREG IVB North West Europe Program) is to develop innovative advisory tools based on the MIR data collected by milk recording organization. One of the first objectives is to develop a tool to assess the pregnancy status of cows. The tool is based on the comparison of the observed spectrum with an expected spectrum obtained from a set of spectra with a known status of the cow, here being open. Development was done using Walloon milk recording data. A training dataset (342,832 spectral data from 66,174 cows) was used to obtain residual spectra (i.e. difference between observed and expected spectra). Based on the fact that the pregnancy status of all cows was known, predictive discriminant function was constructed on 2,154 residual spectra randomly selected from the initial dataset. The discriminant function was then applied on the rest of the dataset (12,160 residual spectra) for validation. When considering the period from 21 to 50 days after an insemination, the error rate was about 1.9% with a specificity of 82.4% and a sensibility of 99.8%. These results showed a high potential for using directly the MIR spectrum of milk to detect a change in the pregnancy status of dairy cows. This methodology can also be applied to predict other types of physiological status changes (e.g. udder health related) and can be used on other types of biomarker data (i.e. collected from on-farm sensors). Similarly, integration of on-farm information on expected pregnancy status could improve the presented off-farm tool.
http://hdl.handle.net/2268/177703

File(s) associated to this reference

Fulltext file(s):

FileCommentaryVersionSizeAccess
Open access
EAAP2014_Absract_ALaine.pdfPublisher postprint181.53 kBView/Open

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.