Indirect selection for feed efficiency using the social animal model: Evaluation by simulation in growing Duroc pigs
Our objective was to predict response on feed conversion ratio (FCR) and economic indexes, after selection for average daily gain (ADG) and backfat thickness (BF) when social animal model (SAM) or animal model (AM) are used for the genetic evaluations. This was conducted simulating a nucleus of 400 sows and 20 boars. Data were generated either with SAM or AM based on variance components sampled in each replicate (1,000) from realizations of the posterior distributions obtained in three-trait analyses with the respective models, using records from 1,144 pigs. When AM was used for genetic evaluation no statistically favourable (probability of negative response lower than 0.9) responses on FCR were predicted. The best situation (-0.03(0.05) ((kg/kg)/generation)) was observed when the same weight was assigned to both traits. In this scenario, the economic response in a market with a penalty for excess of BF was 2.02(1.40) ((€/pig)/generation). Nor were significant responses on FCR observed when SAM was used for evaluation. In this case, the best economic response was also predicted under penalty for excess of BF when ADG and BF had the same weight exclusively on direct genetic effects. When AM was used for data simulation and evaluation, favourable responses on FCR (-0.05(0.02) ((kg/kg)/generation)) were predicted in the scenarios having most of the weight on BF. In these cases, the economic response under penalty for excess of BF reached 2.97(0.69) ((€/pig)/generation). Given our datasets sizes, the high uncertainty of SAM parameters prevents better responses than those observed when AM is assumed and used.