Predicción del peso corporal de guajolotes nativos mexicanos a través de medidas morfométricas

Abstract

The aim of this study was to evaluate the prediction of body weight (BW) of the native Mexican guajolote (NMG) from morphometric measurements (MM) and morphological indices using classification and regression tree analysis (CART). Measures were taken of 244 NMG from the states of Puebla, Chiapas and Campeche. The BW and ten MM were collected, and three morphological indices were estimemated: massiveness (MAS), stockiness (STK) and body condition (CON). The descriptive statistics and Pearson’s correlation (r) of the variables were analyzed and a regression tree was constructed using the CART method. Coefficients of variation <20% were obtained in the MM, an MAS of 13.50%, STK of 111.12% and the CON of 16.81%. Correlations between BW and MM ranged from moderate to high (r = 0.35 to r = 0.91; P < 0.0001). The CON was the variable with the best score (100%) in the normalized importance analysis, followed by MAS (79.2%) and thoracic perimeter (52.8%). The optimal regression tree diagram formed a total of 13 nodes, of which 7 were terminal nodes, demonstrating that the CON is sufficient to predict the BW of NMG. This study allowed to define a prediction model with an observed explained variance of 86.4% and included the CON, body height and wing width, which can be applied by producers to reliably predict body weight of NMG