Paper published in a journal (Scientific congresses and symposiums)
Covariance estimation with Method-R
Druet, Tom; Misztal, I.; Duangjinda, M. et al.
2000In Journal of Animal Science, 78/83 (Suppl. 1/ Suppl. 1), p. 56
Peer Reviewed verified by ORBi
 

Files


Full Text
Covariance estimation with Method-R.pdf
Publisher postprint (206.06 kB)
Request a copy

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
Method-R; covariance; estimation
Abstract :
[en] The objective of the study was to develop algorithms based on MethodR that allow estimation of (co)variance components with large data sets for complex single trait models, e.g., with correlated additive animal effects and/or with dominance effects. Theoretical Method-R formulas were developed for simplified single and bi-variate models. In single trait, the curve of the regression of Method-R was continuous and monotonic, as is described in the literature, and its slope depended on the amount of information on one animal. The curve was flatter as the number of records per animal increased possibly indicating numerical problems with the sire model. For covariance, the curve of the regression was not always monotonic and it had a discontinuity; a regression factor of 1 still corresponded to the correct covariance. Similar curves were observed in analyses of simulated data sets. Due to the observed discontinuity, algorithms implementing Method-R that require continuous regression curve would not work in models with covariances. An alternative algorithm was based on a transformation matrix obtained by multiplying a matrix of numerators with the inverse of a matrix of denominators of the regression factors. This algorithm always converged in models with covariances, but was slow, requiring as many as 1000 rounds to converge. Convergence, faster by 3-10 times, was achieved by applying over-relaxation. Analyses of several simulated and real data sets by Method-R showed that sampling variance of (co)variance estimates with Method-R was higher for covariances or dominance effects than for additive effects. Therefore, larger number of samples is necessary for more complex models to obtain reliable estimates by Method-R.
Disciplines :
Animal production & animal husbandry
Author, co-author :
Druet, Tom ;  Université de Liège - ULiège > Département de productions animales > GIGA-R : Génomique animale
Misztal, I.
Duangjinda, M.
Reverter, A.
Gengler, Nicolas  ;  Université de Liège - ULiège > Agronomie, Bio-ingénierie et Chimie (AgroBioChem) > Ingénierie des productions animales et nutrition
Language :
English
Title :
Covariance estimation with Method-R
Publication date :
July 2000
Event name :
2000 ADSA/ASAS joint annual meeting
Event organizer :
ADSA - ASAS
Event place :
Baltimore, MD, United States
Event date :
July 24–28, 2000
Audience :
International
Journal title :
Journal of Animal Science
ISSN :
0021-8812
eISSN :
1525-3163
Publisher :
American Society of Animal Science, Savoy, United States - Illinois
Special issue title :
Abstracts American Dairy Science Association/ American Society Of Animal Science
Volume :
78/83
Issue :
Suppl. 1/ Suppl. 1
Pages :
56
Peer reviewed :
Peer Reviewed verified by ORBi
Available on ORBi :
since 15 February 2011

Statistics


Number of views
33 (8 by ULiège)
Number of downloads
8 (1 by ULiège)

Bibliography


Similar publications



Contact ORBi