Genetic Determination of Maternal Behaviour in Pigs

In sows, maternal behaviour is a commercially important trait which affects sow and piglet welfare and the quality and safety of human-animal interactions, especially if free farrowing of sows without crates is in place. The aim of this study was to evaluate maternal behaviour and to estimate its quantitative genetic parameters in Swiss Landrace sows (SLR). The lactating sows, with an average lactation period of 29-30 days and different parities (one to eleven) at four nucleus farms of the Swiss pig breeding company Suisag underwent behaviour tests and observations during two 12-week periods. The level of aggressiveness of the sows towards humans was assessed with a standardised Piglet Handling Test and with a Farmers' Aggressiveness Score, using five- and three-point scoring systems, respectively. Additionally, the reaction to the playback of unknown squealing piglets was evaluated using a five-point scoring system, and vocalising before lying down and time taken to lay down were recorded. Depending on the trait, data were collected on 451-705 sows, most of which had phenotypes available for more than one trait. Number of samples per measure varied between one and six. Given the non-Gaussian distribution of the traits, variance components were estimated based on Bayesian generalised linear mixed models, including as random effects the sow genetic effects and the sow permanent environmental effects. The additive genetic relationships were either computed from the pedigree (1418 individuals across six generations), the genomic data (from 376 genotyped sows using SNP-Chip 60 K), or their combination using the H matrix. The credibility intervals of the estimates indicate that all traits were repeatable independently of how the additive genetic relationships were modelled, with estimated repeatabilities varying between 0.18 ± 0.09 (mean ± SD; 95% CI = [0.01, 0.35]) and 0.66 ± 0.04 ([0.58, 0.74]). Both the level of aggressiveness towards humans (pedigree based: Piglet Handling Test: h² = 0.44 ± 0.12, [0.22, 0.69]; Farmers' Aggressiveness Score: h² = 0.24 ± 0.10, [0.04, 0.45]), and vocalising before lying down were found to be heritable (H matrix based: h² = 0.36 ± 0.15, [0.07, 0.63]). We found an unfavourable genetic correlation between the Farmers' Aggressiveness Score and piglet survival traits. Our results suggest that human-directed aggression is heritable, indicating that improvement of stockpersons' safety and sow welfare through genetic selection is feasible, although piglet survival should be included in a balanced breeding index to ensure progress in both traits.