Abstract

SOUZA, J.C. et al. Genetic distances and principal component estimations in brazilian beef cattle. Arch. zootec. [online]. 2010, vol.59, n.228, pp.479-485. ISSN 1885-4494.

The objective of this study was to estimate principal components for, and correlations between, weights at birth (BWT), 205d (WT205), 365d (WT3650), 550d (WT550) and 730d (WT730) using multivariate analysis of data from 2,237 animals from IAPAR, PR. Calves were produced from Nellore, Guzerá, Red Angus, Marchigiana, and Limousin bulls mated to Nellore, ½ Guzerá x ½ Nellore, ½ Red Angus x ½ Nellore, ½ Marchigiana x ½ Nellore, ½ Simmental x ½ Nellore, ¾ Nellore x ¼ Red Angus, and ½ Red Angus x ¼ Guzerá x ¼ Nellore cows. Correlations ranged from 0.22 (BWT with WT730) to 0.73 (WT365 with WT550). The proportions of variation justified by the principal components were: first, 56%; first and second, 72%; first through third, 87%; and first four, 95%. The MANOVA tests model a separate variance for each trait (in this case, a separate variance for each of the five traits, and a separate covariance for each of the pairs of traits). The Wilks' Lambda indicates a significant difference in genetic distance between breeds of dam based on the performance of calves. Dams composed of ½ Marchigiana x ½ Nellore versus ½ Simmental x ½ Nellore had the least genetic distance (0.058), while Nellore versus ½ Simmental x ½ Nellore were most distant (2.738). Multivariate analysis can be used to estimate principal components for developing economic weights and accurate indices important in beef selection. Trees generated from cluster analysis of progeny weights can be used to visualize genetic grouping of breed of dam.

Keywords : Multivariate analysis; Weight.

        · abstract in Portuguese     · text in Portuguese     · Portuguese ( pdf )