Inferring causal relationships among growth curve traits of Lori-Bakhtiari sheep using structural equation models

Abstract

This study aimed to infer the causal effects of birth weight on growth curve traits as well as, assessment of causal relationships among the growth curve traits in Lori-Bakhtiari sheep using structural equation modeling. The body weight-age records were collected from birth to 390 days of age from 1995 to 2012 in Shooli Breeding Station of Lori-Bakhtiari sheep, Shahr-e Kord, Iran. Initially, five non-linear models including Brody, Logistic, Negative exponential, Gompertz and von Bertalanffy were fitted on 22,546 body weight-age records for determining the best model describing the growth curve. Among the tested non-linear growth models, Brody model showed the best goodness of fit. Then, data on growth curve traits of 3,168 lambs from 217 rams and 1,211 ewes were considered for genetic evaluation of growth curve traits including asymptotic weight (A), integration constant related to initial animal weight (b), and maturation rate (k).The causal effect of birth weight (BW) on growth curve traits under two models including standard multivariate model (SMM) and fully recursive model (FRM) were fitted via a Bayesian approach. The posterior means of heritability for the growth curve traits were similar under both SMM and FRM, while the posterior means of genetic correlations among the traits were statistically different under both SMM and FRM. Structural coefficients related to causal effects of BW on b, k and A, were statistically significant with values −0.19, 0.01and 12.58, respectively. Therefore, considering the causal effects of BW on growth curve traits is essential for accurate genetic analysis of growth curve traits of Lori-Bakhtiari sheep.