Genetic parameters for weight gain and body measurements for Nile tilapias by random regression modeling

Abstract

The aim of the current study was to estimate the genetic parameters for weight gain and body measurements in the GIFT (Genetically Improved Farmed Tilapia) strain of Nile tilapia by random regression models. Several orders of Legendre polynomials were tested for random effects and modeled with 1, 3, 6 and 9 classes of residual variance. For the effects of permanent environmental and family, third-order polynomials were adjusted in all traits, as well as for the genetic effects of weight, weight gain, length and width. For genetic effects of height and head, fourth-order polynomials were required. To gain weight, height and head, the best model was one that considered the homogenous residual variance; however for width and weight, heterogeneous variance with 3 and 9 age classes was required, respectively. The highest heritability for weight was 0.34 at 240–311 days, and for weight gain it was 0.69 at 311 days. For head and length, the highest heritability was around 270 days at 0.27 and 0.21, respectively. The highest heritability found for length was 0.20 at 254 days, 0.2 at 254 days for height, and for width the heritability was 0.54 at 311 days. Since the largest heritabilities were found for weight gain and width at 311 days, selection at these ages, based on these traits, would lead to greater genetic gains. Genetic correlations were higher between adjacent ages and, in general, selections at ages of less than 200 days did not lead to genetic gain correlated with traits at 300 days. The exception was for width, because high correlations were obtained between final and initial ages and the heritability was median in the majority of the period. Thus, selection based on the width at any age would lead to satisfactory genetic gain in this trait at the end of the growing season.

Keywords

• Strain; Heritability; Legendre polynomials; Longitudinal data