2015Journal of Dairy ScienceOpen access

Genetic analysis of calving traits by the multi-trait individual animal model

J.I. Weller, E. Ezra

Open full text 16 citations

Abstract

Five alternative models were applied for analysis of dystocia and stillbirth in first and second parities. Models 1 and 2 were included only to estimate the parameters required for model 4, and models 3 and 5 are included only as comparisons to the model 4 estimates. Variance components were estimated by multi-trait REML, including cows with valid calving records for both parities. For the effects of sire of calf on first and second parities, variance components were estimated including only calvings with the same sire of calf for both parities. All heritabilities for the cow effect were quite low, but higher for dystocia than for stillbirth and higher in first parity. The sire-of-calf heritabilities were higher than the cow effect heritabilities, except for stillbirth in parity 2. Unlike the effect of cow correlations, all sire of calf correlations were >0.6, and the correlations for the same trait in parities 1 and 2 were >0.9. Thus, a multi-trait analysis should yield a significant gain in accuracy with respect to the sire of calf effects for bulls not mated to virgin heifers. A multi-trait individual animal model algorithm was developed for joint analysis of dystocia and stillbirth in first and second parities. Relationships matrices were included both for the effects of cow and sire of calf. In addition, random herd-year-season and fixed sex of calf effects were included in the model. Records were preadjusted for calving month and age. A total of 899,223 Israeli Holstein cows with first calvings since 1985 were included in the complete analysis. Approximate reliabilities were computed for both sire of cow and sire of calf effects. Correlations between these reliabilities and reliabilities obtained by direct inversion of the coefficient matrix for a sire of cow-sire of calf model were all close to 0.99. Phenotypic trends for cows born from 1983 through 2007 were economically unfavorable for dystocia and favorable for stillbirth in both parities. Genetic trends were economically unfavorable for both dystocia and stillbirth in first parity. First-parity sire of calf trends were unfavorable for dystocia, but favorable for stillbirth. All environmental trends were nearly zero. Regressions of evaluations of the complete analysis on a model including only calvings before 2011 were all >0.8. All evaluations met the Interbull Method 3 criterion for unbiasedness. Model 4, which computed genetic evaluations for all animals for all 4 traits accounting for all known relationships and correlations among the traits, is recommended for routine genetic evaluation of calving traits.

Open-access reader

About this research paper

What this paper is about

Five alternative models were applied for analysis of dystocia and stillbirth in first and second parities. Models 1 and 2 were included only to estimate the parameters required for model 4, and models 3 and 5 are included only as comparisons to the model 4 estimates. Variance components were estimated by multi-trait REML, including cows with valid calving records for both parities. For the effects of sire of calf on first and second parities, variance components were estimated including only calvings with the same sire of calf for both parities. All heritabilities for the cow effect were quite low, but higher for dystocia than for stillbirth and higher in first parity. The sire-of-calf heritabilities were higher than the cow effect heritabilities, except for stillbirth in parity 2. Unlike the effect of cow correlations, all sire of calf correlations were >0.6, and the correlations for the same trait in parities 1 and 2 were >0.9. Thus, a multi-trait analysis should yield a significant gain in accuracy with respect to the sire of calf effects for bulls not mated to virgin heifers. A multi-trait individual animal model algorithm was developed for joint analysis of dystocia and stillbirth in first and second parities. Relationships matrices were included both for the effects of cow and sire of calf. In addition, random herd-year-season and fixed sex of calf effects were included in the model. Records were preadjusted for calving month and age. A total of 899,223 Israeli Holstein cows with first calvings since 1985 were included in the complete analysis. Approximate reliabilities were computed for both sire of cow and sire of calf effects. Correlations between these reliabilities and reliabilities obtained by direct inversion of the coefficient matrix for a sire of cow-sire of calf model were all close to 0.99. Phenotypic trends for cows born from 1983 through 2007 were economically unfavorable for dystocia and favorable for stillbirth in both parities. Genetic trends were economically unfavorable for both dystocia and stillbirth in first parity. First-parity sire of calf trends were unfavorable for dystocia, but favorable for stillbirth. All environmental trends were nearly zero. Regressions of evaluations of the complete analysis on a model including only calvings before 2011 were all >0.8. All evaluations met the Interbull Method 3 criterion for unbiasedness. Model 4, which computed genetic evaluations for all animals for all 4 traits accounting for all known relationships and correlations among the traits, is recommended for routine genetic evaluation of calving traits.

Why it matters

OpenAlex reports 16 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Five alternative models were applied for analysis of dystocia and stillbirth in first and second parities. Models 1 and 2 were included only to estimate the parameters required for model 4, and models 3 and 5 are included only as comparisons to the model 4 estimates. Variance components were estimated by multi-trait REML, including cows with valid calving records for both parities. For the effects of sire of calf on first and second parities, variance components were estimated including only calvings with the same sire of calf for both parities. All heritabilities for the cow effect were quite low, but higher for dystocia than for stillbirth and higher in first parity. The sire-of-calf heritabilities were higher than the cow effect heritabilities, except for stillbirth in parity 2. Unlike the effect of cow correlations, all sire of calf correlations were >0.6, and the correlations for the same trait in parities 1 and 2 were >0.9. Thus, a multi-trait analysis should yield a significant gain in accuracy with respect to the sire of calf effects for bulls not mated to virgin heifers. A multi-trait individual animal model algorithm was developed for joint analysis of dystocia and stillbirth in first and second parities. Relationships matrices were included both for the effects of cow and sire of calf. In addition, random herd-year-season and fixed sex of calf effects were included in the model. Records were preadjusted for calving month and age. A total of 899,223 Israeli Holstein cows with first calvings since 1985 were included in the complete analysis. Approximate reliabilities were computed for both sire of cow and sire of calf effects. Correlations between these reliabilities and reliabilities obtained by direct inversion of the coefficient matrix for a sire of cow-sire of calf model were all close to 0.99. Phenotypic trends for cows born from 1983 through 2007 were economically unfavorable for dystocia and favorable for stillbirth in both parities. Genetic trends were economically unfavorable for both dystocia and stillbirth in first parity. First-parity sire of calf trends were unfavorable for dystocia, but favorable for stillbirth. All environmental trends were nearly zero. Regressions of evaluations of the complete analysis on a model including only calvings before 2011 were all >0.8. All evaluations met the Interbull Method 3 criterion for unbiasedness. Model 4, which computed genetic evaluations for all animals for all 4 traits accounting for all known relationships and correlations among the traits, is recommended for routine genetic evaluation of calving traits.

Key concepts: Sire, Ice calving, Trait, Herd, Parity (physics), Restricted maximum likelihood, Animal science, Random effects model

Related papers

Back to paper searchBrowse research topicsOriginal source
Genetic analysis of calving traits by the multi-trait individual animal model — Research Paper | ScholarLens