Selection of variables for predicting and comparing number of skin lesions

Abstract Text:

Legislative and consumer pressure to improve sow welfare is driving the transition to group housed systems for gestating sows. While group housing provides increased space and socialization, there is increased potential for aggression that can cause injury and stress leading to poor production and increased culling rates. This work is part of a larger study aimed at estimating genetic and environmental sources of variation affecting aggression in group housed prepubertal and gestating sows. The objective of this work is to estimate systematic environmental effects affecting the number of skin lesions, which are a proxy for inferring levels and types of aggressive interactions between pigs. 320 purebred Yorkshire pigs are being housed in groups of 10 pigs per pen (0.21 m²/pig) in nursery pens and 13-15 pigs per pen (0.84 m²/pig) in finisher and gilt pens. Remixing by weight to emulate mixing in commercial settings occurs at weaning, at transition to grow-finish, and at gilt retention for breeding. At each mixing the number of familiar and unfamiliar pigs is held constant for each group. Numbers of fresh lesions (< 24 hours old) were collected pre-mixing, 24-hours post-mixing, and 3 weeks post-mixing to examine individual response to mixing and performance in established social settings. Analysis was performed using a generalized over-dispersed Poisson mixed model. Fixed effects included phase of production (3 levels), observer (2 levels), sex (2 levels), measure (3 levels), and a live weight covariate. All fixed effects were deemed significant (P < 0.0001). Both nursery pigs and market weight gilts had significant differences in total lesions prior to mixing and 3 weeks post-mixing, as well as between 24-hours post-mix and 3 weeks post-mix (P < 0.01). There was no significant difference between pre-mixing and 3 weeks post-mixing lesion numbers suggesting that pigs return to pre-mixing levels of aggression (Nursery: P = 0.8479, Sow: P = 0.1865). Finishing pigs had significant (P < 0.0001) differences in total number of lesions between each measure, however the least square mean differences between pre-mix and 3 weeks post-mixing were minor (0.7404 compared to 1.1608 and 1.9012). These preliminary results are encouraging because they suggest that there is extensive variation in skin lesions that can be modeled and attributed to systematic factors. Given the variation observed in the phenotype, it is recommended to increase the sample size to 1120 pigs across 8 replications as planned in future stages of this study.

Keywords:

Pigs, group housing, lesions