Exploring the Use of Random Regression Models withLegendre Polynomials to Analyze Clutch Sizein Iranian Native Fowl

Random regression models (RRM) have become common for the analysis of longitudinal data or repeated records on individual over time. The goal of this paper was to explore the use of random regression models with orthogonal / Legendre polynomials (RRL) to analyze new repeated measures called clutch size (CS) as a meristic trait for Iranian native fowl. Legendre polynomial functions of increasing order 0 (no covariate) to 4 were fitted to the age at sexual maturity (ASM) and 1 to 10 to the additive genetic and permanent environmental effects. Days in production (clutch) were used as time variables. Homogeneity of residual variance through the time was assumed. Analyses were carried out within restricted maximum likelihood algorithm (REML) using WOMBAT software. Adequacy of models was checked by Bayesian information criterion (BIC). The resulted BICs suggested a model composed of the second order polynomial for ASM and 8th order polynomial for additive genetic and permanent environmental effect was the most suitable for adjusting the present records. The highest phenotypic and permanent environmental variance of CS was at the beginning of the production period. Additive genetic variance was fairly consistent during 210 and 265 days of age (d 210-d265). Estimates of heritability for CS ranged from 0.033 to 0.199 for d 161 and d 242 in the first cycle of egg production, respectively. The ratio of animal permanent environmental variance to phenotypic variance was in the range of 0.01 and 0.264. The estimated ranges for additive genetic and permanent environmental correlations were -0.18 to 0.99 and -0.5 to 0.99, respectively and were high between the adjacent ages and they tended to decrease at nonadjacent ages.