HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text Full Text (PDF) Interpretive Summary Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Google Scholar Articles by Holtsmark, M. Articles by Ødegård, J. PubMed PubMed Citation Articles by Holtsmark, M. Articles by Ødegård, J.

Predictive abilities of different statistical models for analysis of survival data in dairy cattle

M. Holtsmark^*,,1, B. Heringstad^*, and J. Ødegård^*,

^* Department of Animal and Aquacultural Sciences, Norwegian University of Life Sciences, PO Box 5003, N-1432 Ås, Norway
Geno Breeding and A.I. Association, PO Box 5003, N-1432 Ås, Norway
Nofima, PO Box 5010, N-1432 Ås, Norway

¹ Corresponding author: marte.holtsmark{at}umb.no

The objective of this study was to compare alternative trait definitions and statistical models for genetic evaluation of survival in dairy cattle. Data from the first 5 lactations of 808,750 first-crop daughters of 3,064 Norwegian Red sires were analyzed. Seven sire models were used for genetic analyses: linear and threshold cross-sectional models for binary survival scores from first lactation; a linear multi-trait model for survival scores from the first 3 lactations; linear and threshold repeatability models for survival scores from the first 5 lactations; a Weibull frailty model for herd life in first lactation; and a Weibull frailty model for herd life in the first 5 lactations. The models were compared to assess predictive ability of sire estimated breeding values with respect to average survival 365 d after first calving for second-crop daughters (not included in calculation of predicted transmitting abilities) of 375 elite sires. Generally, the linear multi-trait model analyzing survival in the first 3 lactations as correlated traits gave more-accurate predicted sire breeding values compared with both linear and Weibull frailty models using data from first lactation only, even when the latter models were extended to include data up to the sixth lactation. The Weibull frailty models did not improve predictive ability of sire estimated breeding values over what was obtained using a simple cross-sectional linear model for binary survival in first lactation.

Key Words: predictive ability • model comparison • survival