The aim of this study was to elucidate the structure of relationships between milk yield, composition, and coagulation properties of Brown Swiss cattle. Multi- variate factor analysis was used to derive new synthetic variables that can be used for selection purposes. For this reason, genetic parameters of these new variables were estimated. Individual records on milk yield, fat and protein percentages, casein content, lactose per- centage, somatic cell count, titratable acidity, and pH were taken on 1,200 Italian Brown Swiss cows located in 38 herds. Factor analysis was able to extract 4 latent variables with an associated communality equal to 70% of the total original variance. The 4 latent factors were interpreted as indicators of milk composition, coagula- tion, acidity, and mammary gland health, respectively. Factor scores calculated for each animal exhibited coherent patterns along the lactation and across dif- ferent parities. Estimation of genetic parameters of factor scores carried out with a multiple-trait Bayesian hierarchical model showed moderate to low heritabili- ties (raging from 0.10 to 0.23) and genetic correlations (from −0.15 to 0.46). Results of the present study sug- gest the hypothesis of a simpler structure that controls, at least in part, the covariance of milk composition and coagulation properties. Moreover, extracted variables may be useful for both breeding and management pur- poses, being able to represent, with a single value for each animal, complex traits such as milk coagulation properties or health status of the mammary gland.
Use of multivariate factor analysis to define new indicator variables for milk composition and coagulation properties in Brown Swiss cows
MELE, MARCELLO;
2012-01-01
Abstract
The aim of this study was to elucidate the structure of relationships between milk yield, composition, and coagulation properties of Brown Swiss cattle. Multi- variate factor analysis was used to derive new synthetic variables that can be used for selection purposes. For this reason, genetic parameters of these new variables were estimated. Individual records on milk yield, fat and protein percentages, casein content, lactose per- centage, somatic cell count, titratable acidity, and pH were taken on 1,200 Italian Brown Swiss cows located in 38 herds. Factor analysis was able to extract 4 latent variables with an associated communality equal to 70% of the total original variance. The 4 latent factors were interpreted as indicators of milk composition, coagula- tion, acidity, and mammary gland health, respectively. Factor scores calculated for each animal exhibited coherent patterns along the lactation and across dif- ferent parities. Estimation of genetic parameters of factor scores carried out with a multiple-trait Bayesian hierarchical model showed moderate to low heritabili- ties (raging from 0.10 to 0.23) and genetic correlations (from −0.15 to 0.46). Results of the present study sug- gest the hypothesis of a simpler structure that controls, at least in part, the covariance of milk composition and coagulation properties. Moreover, extracted variables may be useful for both breeding and management pur- poses, being able to represent, with a single value for each animal, complex traits such as milk coagulation properties or health status of the mammary gland.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
