Multivariate approach in the evaluation of performance and carcass traits of Suffolk crossbred lambs

Abstract

Evaluate associations between economically important traits is crucial as it considers the intensity of therelationship between variables and aids in excluding redundant traits, thereby facilitating the early selection of animals.The aim of this study was to evaluate the association between performance and carcass traits of crossbred lambs and to identifya trait that demonstrates higher discriminatory power, to assisting in the selection of animals. Was used 61 male lambs wereused to assess trait associations via Pearson correlation, Euclidean distance, and principal components (PC). With clusteranalysis, we observed the formation of two distinct groups (Euclidean distance), indicating significant dissimilarity betweenthe groups. This dissimilarity was attributed to the group of variables RP, TD, TW, and RW, while the other group wascharacterized by LP, SW, HCW, and ECC. It was observed that performance with carcass traits presented linear correlationsbetween 0.28 to 0.63. Through multivariate analysis, it was possible to 4 PCs selected that together explained 78.35% of thetotal variance of the data, with eigenvalues greater than 0.70. Through the first principal component, which retains the highestpercentage of total variance (38.94%), redundant or non-discriminant descriptors were discarded. Within this component,four traits (SW, ECC, RW, and HCW) were considered the most important in describing the variability of the dataset studied.However, slaughter weight would be the selectable variable to represent the other carcass traits, due to greater discriminatorypower. This trait can be considered easier to measure and more economical for breeders.