Estimation of genetic parameters for litter size in swine by crossbred model

Abstract

The objectives of this study were to estimate genetic parameters, genetic correlation between purebred and crossbred, and compare accuracy and rank correlation of EBV using models from several datasets: 1) within breed model using purebred data set (Model I), 2) pooled data set using animal model (Model II) and 3) Pooled data set using crossbred model (Model III). The data for total born (TB) and born alive (BA) of 63,284 records of purebred Large White (LW), Landrace (LR) and two-breed cross (TBC) from a commercial swine company during year 1998 to 2003 were used in the analysis. The variance components and genetic parameters were estimated by Restricted Maximum Likelihood (REML) using BLUPF90 PigPAK 2.5. The results found that heritabilities were analyzed by Model I, II, and III not different. The genetic correlation between purebreds and crossbreds (rpc) ranged from 0.20-0.30. The rpc was less than 0.8 suggesting that a possible benefit from combined evaluation purebreds and crossbreds. This study showed that model II had the lowest accuracy in BV estimates, while the estimated breeding value (EBV) from model III had highest accuracy. The rank correlation between EBV of 50 top sires and dams from model I and model III ranged from 0.920 and 0.982, which showed that ranks of sire and dam of both models were similar.