Genotype-by-environment (GxE) interactions are sometimes of sizeable magnitude and can consequently affect significantly the accuracy of genetic and phenotypic predictions. This is especially relevant under climate change and for dairy cattle, given the worldwide range of breeding populations. Several methods have been proposed to account for GxE interactions in genetic models, from multiple-trait models to reaction-norms. In this work, we propose to use the enviromics approach, which is based on the measuring of environmental covariables (envirotyping) that provide an environmental relatedness matrix: together with the kinship matrix, this approach can better capture the covariance between records and GxE interactions compared to the modelling of environments as all-encompassing black boxes. As a first preliminary step, we estimated the size of GxE interactions in Holstein-Friesian dairy cattle. We took the average milk production records of the daughters of 611 sires from three territories in Northern Italy, in the year 2021: Cremona, Mantova and Cuneo. The kinship matrix was estimated from pedigree records. Days in milk, the average age of cows, the month and day of milk recording and the number of days from the latest insemination were used as systematic effects. A linear mixed model that treated milk from each environment as a separate trait was fitted to the data. The MCMC method implemented in the BGLR R package was used to solve the model. Heritability estimates ranged from 0.43 to 0.49, while genetic correlations ranged from 0.367 to 0.414. These results indicate the presence of significant GxE interactions that should be accounted for in genetic models. The collating of environmental descriptors is currently in progress and, when completed, it will allow us to construct the environmental covariance matrix. This enviromics approach will be benchmarked against traditional strategies for the modelling of GxE interactions in terms of i) the estimation of breeding values and their ranking across environments, and ii) the accuracy of phenotypic predictions across environments.
An enviromics approach to modelling GxE interactions in dairy cattle
Roberta CiampoliniMembro del Collaboration Group
;
2025-01-01
Abstract
Genotype-by-environment (GxE) interactions are sometimes of sizeable magnitude and can consequently affect significantly the accuracy of genetic and phenotypic predictions. This is especially relevant under climate change and for dairy cattle, given the worldwide range of breeding populations. Several methods have been proposed to account for GxE interactions in genetic models, from multiple-trait models to reaction-norms. In this work, we propose to use the enviromics approach, which is based on the measuring of environmental covariables (envirotyping) that provide an environmental relatedness matrix: together with the kinship matrix, this approach can better capture the covariance between records and GxE interactions compared to the modelling of environments as all-encompassing black boxes. As a first preliminary step, we estimated the size of GxE interactions in Holstein-Friesian dairy cattle. We took the average milk production records of the daughters of 611 sires from three territories in Northern Italy, in the year 2021: Cremona, Mantova and Cuneo. The kinship matrix was estimated from pedigree records. Days in milk, the average age of cows, the month and day of milk recording and the number of days from the latest insemination were used as systematic effects. A linear mixed model that treated milk from each environment as a separate trait was fitted to the data. The MCMC method implemented in the BGLR R package was used to solve the model. Heritability estimates ranged from 0.43 to 0.49, while genetic correlations ranged from 0.367 to 0.414. These results indicate the presence of significant GxE interactions that should be accounted for in genetic models. The collating of environmental descriptors is currently in progress and, when completed, it will allow us to construct the environmental covariance matrix. This enviromics approach will be benchmarked against traditional strategies for the modelling of GxE interactions in terms of i) the estimation of breeding values and their ranking across environments, and ii) the accuracy of phenotypic predictions across environments.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
