Genetic relationships between temperament traits of sows during pregnancy and lactation

Group-housed gestating sows face challenging situations, including contact with the stockpersons and aggressive interactions with their pen-mates. During lactation, husbandry procedures may trigger maternal defensive aggression in sows. This study explored the genetic determination of the response to humans and the number of skin lesions resulting from social aggression in gestating Swiss Landrace sows and to estimate their correlations with maternal defensive aggression and piglet survival. During two 12-week periods, gestating sows (parities one to eleven) at four nucleus farms of the breeding company Suisag underwent a Forced Human Approach Test (n = 489). Additionally, whether they approached the experimenter was recorded. The number of anterior (n = 356) and posterior (n = 349) fresh skin lesions from social aggression were counted. Sows tested during gestation were also assessed during lactation (n = 388) using a Piglet Handling test to evaluate maternal defensive aggression. Of these, 334 sows had their level of aggressiveness during lactation scored by farmers. The number of repetitions per measure varied between one and eight. Variance components were estimated using Bayesian analysis based on animal models, including as random effects the sow's permanent environment and genetic or genomic effect. The additive genetic relationships were modelled using pedigree data (1 418 individuals, six generations), genomic data (365 genotyped sows, 60K SNP-Chip), or their combination. The Forced Human Approach Test, whether the sow approached the human, and the number of anterior and posterior skin lesions were all temporally stable, with repeatabilities ranging from 0.36 ± 0.09 to 0.63 ± 0.10 (mean ± SD). While the heritability of approaching the human was lower than 0.10, corresponding estimates for the other behavioural traits ranged from 0.15 ± 0.16 to 0.33 ± 0.17. The negative genetic correlations between the level of tolerance to humans during gestation and maternal defensive aggression during the following lactation (r = -0.36 ± 0.19) indicate that more tolerant sows were less likely to be aggressive. However, a higher level of tolerance was genetically associated with lower piglet survival (number of piglets' deaths due to crushing: r = 0.51 ± 0.23; proportion of piglets survived to weaning: r = -0.32 ± 0.39). Previously, we estimated unfavourable genetic correlations between farmers' scores of aggressiveness and piglet survival. As the response to a human of gestating sows is favourably genetically correlated with the level of maternal defensive aggressiveness when the piglets are handled, genetic selection for calmer sows could be implemented using temperament traits measured during gestation. However, this might have a moderate negative impact on piglet survival unless balanced in a multitrait breeding index.