بررسی تاثیر تعداد سطوح یک صفت آستانه‌ای بر دقت برآورد ارزش های ارثی، روندهای ژنتیکی، فنوتیپی و همخونی یک جمعیت اصلاحی

To investigate the effects of categorical trait levels on EBVs accuracies, genetic and phenotypic trends and inbreeding coefficient, a breeding stock contained 400 animals was simulated through stochastic method using R programming package. The population was simulated for 7 generations while each generation contained 200 animals. The generations were overlapped and in each ones 20 percent of the best males and all of the females had a chance to participate in the mating program up to 5 generations. A categorical trait with heritability of 0.25, mean of 5 and standard deviation of 2 was simulated, and on the basis of recording approach, 4 scenarios were designed. In the first scenario a threshold trait was measured as continuous records, while for scenarios 2 to 4, regarding to the appraiser ability a threshold trait were recorded in 3, 5 and 9 categories respectively. Breeding values were estimated by a simple, single trait animal model and all scenarios were replicated for 10 times. The accuracy of breeding value estimations for scenarios 1 to 4 were 0.62, 0.49, 0.59 and 0.61 respectively (p<0.01). Through 7 generations, the genetic trends for scenarios 1 to 4 were 0.212, 0.162, 0.195 and 0.205 and the phenotypic trends were 0.185, 0.137, 0.163 and 0.179 respectively (p<0.01) .The mean of inbreeding coefficient for scenarios 1 to 4 were estimated 0.0053, 0.0052, 0.0052 and 0.0050 respectively (p>0.05). Results showed that, the improvement of recording accuracies for threshold traits lead to increasing in the accuracy of breeding value estimations and the genetic gain. So this is concluded that the development of more accurate measurement tools for threshold traits, could improve the genetic gain and breeding program efficiency.