Selection of traits in poultry breeding using cluster analysis

Abstract

This study aimed to select representative traits from phenotypic variability between broiler chicken grandparent lines using cluster analysis. Chicks from four dam lines and two sire lines of broiler grandparents were housed to assess performance traits (body weight, eviscerated carcass yield, breast yield and leg yield) and morphometric traits (back and leg feathering), which were used to calculate the Mahalanobis' distance, between chicken line pairwise. Phenograms were obtained based on those distances using UPGMA method in cluster analysis. The smallest relative contribution of traits for divergence between chicken lines was assessed to eliminate them. Body weight at 35 days, all carcass traits and back feathering at 28 days were the most representative traits to explain phenotypic variability between chicken lines. From observed clusters and the largest heterogeneity between chicken lines from different clusters, crosses between sire lines SL1 and SL2 with dam lines DL3 and DL4 might result in strain combinations with the largest heterosis. Cluster analysis on performance and morphometric traits is suitable to study phenotypic variability between broiler chicken grandparent lines. We suggest analyzing data from poultry breeding using cluster analysis. © Asian Network for Scientific Information, 2008.