Random regression models for daily feed intake in Danish Duroc pigs

Publikation: KonferencebidragKonferenceabstrakt til konferenceForskning

  • Anders Bjerring Strathe, 1Department of Clinical Veterinary and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Danmark
  • Thomas Mark, 1Department of Clinical Veterinary and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Danmark
  • Just Jensen
  • Bjarne Nielsen, Danmark
  • Do Ngoc Duy, 1Department of Clinical Veterinary and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Danmark
  • Haja N Kadarmideen, 1Department of Clinical Veterinary and Animal Sciences, Faculty of Health and Medical Sciences, University of Copenhagen, Danmark
The objective of this study was to develop random regression models and estimate covariance functions for daily feed intake (DFI) in Danish Duroc pigs. A total of 476201 DFI records were available on 6542 Duroc boars between 70 to 160 days of age. The data originated from the National test station and were collected using ACEMO electronic feeders in the period of 2008 to 2011. The pedigree was traced back to 1995 and included 17222 animals. The phenotypic feed intake curve was decomposed into a fixed curve, being specific to the barn-year-season effect and curves associated with the random pen-year-season, permanent, and animal genetic effects. The functional form was based on Legendre polynomials. A total of 64 models for random regressions were initially ranked by BIC to identify the approximate order for the Legendre polynomials using AI-REML. The parsimonious model included Legendre polynomials of 2nd order for genetic and permanent environmental curves and a heterogeneous residual variance, allowing the daily residual variance to change along the age trajectory due to scale effects. The parameters of the model were estimated in a Bayesian framework, using the RJMC module of the DMU package, where weakly informative priors were derived from preliminary AI-REML analysis. We used a burn-in of 10k followed by 90k rounds, interleaving every 10th sample. Posterior mean heritabilities were low for DFI, but increased from 0.07 to 0.12 at 100 days of age and then decreased fluctuating to around 0.08. The low heritabilities were due to a higher residual variance when using individual DFI records compared to average DFI records over the entire test period. The heritability of DFI for the entire test period was 0.28, corresponding closely to the estimate derived from analysis of average DFI records. Eigenvalues of the genetic covariance function showed that 33% of genetic variability was explained by the individual genetic curve of the pigs. This proportion was covered by linear (27%) and quadratic (6%) coefficients. Genetic eigenfunctions revealed that altering the shape of the feed intake curve by selection may be an option in Danish Duroc pigs
OriginalsprogEngelsk
Udgivelsesår27 okt. 2013
Antal sider1
StatusUdgivet - 27 okt. 2013
Begivenhed The American Association for Laboratory Animal Science National Meeting - Baltimore Convention Center, Minneapolis, USA
Varighed: 27 okt. 201331 okt. 2013

Konference

Konference The American Association for Laboratory Animal Science National Meeting
LokationBaltimore Convention Center
LandUSA
ByMinneapolis
Periode27/10/201331/10/2013

Se relationer på Aarhus Universitet Citationsformater

ID: 52846672