Investigating the Presence of Social Effects in a Swine Nucleus Herd

Interactions between individuals penned together, known as social effects, have been shown to impact growth rate due to competition for food and space. It is not typical for social effects to be used in selection decisions within breeding programs. However, if these effects improve predictability of performance they should be considered. The goal of this study was to determine if social effects can be estimated and used to make selection decisions within a swine breeding program. Two traits were examined, average daily gain from 11 to 23 weeks of age (ADGF) and average daily feed intake (ADFI) for this same period, for three pure breeds, Yorkshire, Landrace, and Duroc. For ADGF, there were 38,854 Yorkshire, 44,365 Landrace, and 59,860 Duroc pigs with records. For ADFI, there were 5,670 Yorkshire, 5,485 Landrace, and 12,194 Duroc records. ADGF is measured on gilts and boars on 7 farms, whereas ADFI is only collected on boars sent to a specialized facility equipped with feed intake recording equipment. All pigs had ad libitum access to feed. Model 1 for ADGF included age deviation from 77 days at on-test as a covariate, sex and barn as fixed effects, pen and litter as random effects, and an additive genetic effect for the individual. Model 1 for ADFI included all the same terms as ADGF with the exception of sex. An additive genetic effect for pen-mates was added to each model in order to estimate social effects (Model 2). Variance estimates for the social effect were low for all breeds and all traits, being less than 1% of the estimate for additive genetic variance. Negative correlation existed between direct and social effects for ADGF in all breeds (values greater than -0.1) and ADFI in Landrace (-0.30). Positive correlation between direct and social effects existed for ADFI in Yorkshire and Duroc (0.40 and 0.005, respectively). The total heritable variance was higher for Model 2 versus Model 1 for all breeds and traits except for ADFI for Landrace which had a lower total heritable variance for Model 2. Rank correlations between the breeding values for the additive genetic effect of the Model 1 and 2 was greater than 0.999 (P<.0001). Based on these results, there is evidence that including social effects could impact selection decisions and further investigation is required to determine how best to implement in a selection program.