Random regression models to explore genetic variation and genetic variability in the growth curve of Baluchi lambs

Abstract

The aim was to apply Random Regression Model (RRM) to describe growth curve of Baluchi lambs. Data was retrieved from an experimental population of Baluchi sheep and consisted of body weight records from 50 to 400 days of age. Regressions for the direct and maternal random effects in the RRM were modeled using different combinations of legendre polynomials with orders from one (i.e., simple repeatability model, SRM) to four. Mean trends were also modeled through a quadratic regression on orthogonal polynomials of age. Homogeneity and heterogeneity of error variance were considered along growth trajectory. The Akaike's Information Criterion (AIC) and Bayesian Information Criterion (BIC) were computed to compare different models. According to AIC and BIC values, simple repeatability model was the worse model and resulted in greater estimates of error variance. A RRM with legendre polynomials of orders 4, 4, 4, and 4 for direct additive genetic, direct permanent environment, maternal additive genetic and maternal permanent environmental effects was selected as the parsimonious model. The AIC and BIC values significantly increased when maternal effects excluded from the parsimonious model. Direct heritability (h2) decreased from 0.32 at 50 days of age to 0.09 at 90 days of age and increased thereafter to a peak at 400 days of age (0.42). Direct permanent environmental effect (p2) was minimum at 50 days of age (0.20), increased gradually and reached the maximized value around 150 days of age (0.54) and decreased thereafter. Maternal heritability (m2) was close to zero at 50 days of age, but increased thereafter to a peak at 400 days of age (0.08). Maternal permanent environmental effect (c2) was 0.22 at 50 days of age, decreased sharply to 0.01 at 100 days of age when it fixed below 0.05 until end of trajectory. Measures of genetic and environmental variability (coefficient of variations: CVA, CVP, CVM and CVC) had the almost similar patterns with corresponding genetic and environmental parameters. Direct and maternal correlations were positively high between adjacent weighs but decreased with increasing number of days between records. In general, estimates of genetic parameters and coefficient of variations showed opportunity for genetic improvement in growth curve of Baluchi lambs.