Differentiation of Exotic Chicken Breeds based on Egg Quality Traits using Classification Tree Algorithm

Abstract

This study aimed to evaluate the effect of breed on egg quality traits and to discriminate chicken breeds (i.e., Bovan Brouwn (BV), Fayoumi (FM) and  Sasso (SS)) based on egg quality traits using classification tree algorithm (CTA). For this purpose, a total of 1696 eggs were utilized and the traits egg 
weight, albumen height, albumen weight, egg length, egg width, egg weight, shell weight, shell thickness, yolk colour, yolk height and yolk weight, were recorded. Significant differences (P<0.05) were found in egg weight, albumen height, albumen weight, and egg width, with Bovans exhibiting the highest LSM values, indicating superior egg quality. Conversely, Sasso demonstrated superior yolk and shell traits with the highest LSM values for yolk weight, yolk height, shell weight, and egg length. Yolk colour differed significantly, with Sasso displaying the darkest yolks. No significant differences were observed in shell thickness. The CTA effectively discriminated between breeds based on egg quality traits, with the model explaining 90% of the variance (R² = 0.90) in the training dataset and 76% (R² = 0.76) in the validation dataset. Key predictors identified included egg weight, albumen height, yolk weight, albumen weight, and egg length. The model achieved high classification accuracy, with ROC AUC values of 0.99 for Bovans, 1.00 for Fayoumi, and 0.98 for Sasso, underscoring its robustness. These findings highlight the critical role of specific egg quality traits in breed differentiation and suggest that CTA is a valuable tool for breeders to enhance productivity and quality by selecting for desirable traits. Further research into the genetic underpinnings of these traits could refine classification models and improve breeding outcomes.