Estudo de interação genótipo x ambiente para peso e perímetro escrotal à desmama na raça Canchim
Resumen
In beef cattle genetic evaluation programs, the existence of genotype x environment
interaction is of special interest when merit of the animals depends on the environment they
are evaluated, that is, when a genotype is superior in one environment but not in others. The
objective in this study was to evaluate the existence of genotype x environment interaction for
weight (WW) and scrotal circumference (SC) at weaning in Canchim cattle, using data
belonging to herds from Southeast and Mid West regions of Brazil. Herds were grouped
according to two criteria: (1) by geographical region, considering animals born in Southeast
and Mid West regions of Brazil; (2) by homogeneous production regions, considering only
animals born in the State of São Paulo. Data on 29,367 WW and 8,800 SC were used for
analyses in criterion 1, and 15,041 WW for analyses in criterion 2. The region in which the
calf was born and raised was considered as the environment, and two methods were used to
evaluate the interaction by both criteria. In method 1, genetic correlation between the same
trait among the environments, in two (criterion 1) and three-trait (criterion 2) analyses, was
estimated, considering the trait in each region as a different trait, using the derivative free
restricted maximum likelihood method, with a statistical model which included the fixed
effects of contemporary group (breeder-owner-year and season of birth-sex-genetic group of
dam-feeding regime) and the covariates age of dam (linear and quadratic effects) and age of
calf (linear effect; only for SC), and the additive direct and maternal, permanent
environmental and residual random effects. Spearman rank correlations for breeding values of
bulls in different regions were also calculated. In method 2, for both criteria, one-trait
analyses using two similar models were done, but one with and the other without the
uncorrelated sire-region random effect, and the difference between them was tested by the
likelihood ratio test. In criterion 1, the genetic correlations between WW and SC in Southeast
and Mid West regions were equal to 0.78 and 0.97, respectively, while in criterion 2, the
genetic correlations between WW in West of São Paulo - Paraná and Araraquara regions,
West of São Paulo - Paraná and Leiteira regions, and Araraquara and Leiteira regions were
0.87, 0.69 and 0.76, respectively. The rank of bulls according to their breeding values
changed from one region to the others, and the models with and without sire-region effect
were different, for all traits. These results suggest the existence of genotype x environment
interaction for the traits in the studied population.