New breeding strategies for better longevity and animal welfare in future crossbred sows

Crossbred sows, which are the result of crossing two purebreds of different breeds, are the dams of the slaughter pigs, and therefore, are economically the most important animals in the pork production chain. Crossbred sows also tend to produce larger litters and have better mothering ability than purebred sows. Being able to produce many high-quality litters without reduction in health and welfare, also known as longevity, is very important in pig production. However, because the true longevity of a sow can only be recorded late in life, it is a difficult trait to select for in purebred animals because they are replaced on breeding values before their true longevity can be observed. An alternative is to collect data on longevity in crossbred sows, which are typically kept longer in production. Using crossbred data to improve purebreds does create some statistical challenges due to differences in genetic background and environment, so these need to be accounted for. The aim of this study was to improve longevity in crossbred sows, as it is important for Norwegian swine genetics for the future. With the use of sensor technology, the goal was to improve the prediction of longevity in crossbred sows further than what is possible with traditional methods. Individual temperature of the sows was measured with a thermographic camera over time during late gestation, farrowing and weaning. Ammonia was measured in the environment using a sensor to look at effects of ammonia on sow performance and longevity. Both the images from the camera and the results from the ammonia sensor created a new type of data structure. Therefore, machine learning tools were also applied to analyse the new data, as well as existing data. The results so far indicate that weight and fat depth at different time points in a sow’s life are highly heritable. Additionally, we saw clear trends for how these traits develop throughout the life of the sows. There are also genetic correlations between these traits and longevity of sows. For the temperature measurements, we saw that temperature in the first couple of days after farrowing was heritable. However, neither traditional nor machine learning methods found any clear relationship between temperature and longevity, or temperature and other traits. There are still additional analyses to be done that could result in different conclusions regarding temperature, however. For the ammonia measurements, the sensor that was initially used was not working optimally and resulted in sub-standard data. What can be said is that ammonia levels in the herds used in this study were low and would likely not impact sows negatively. It may be interesting to repeat the measurements in herds with higher levels of ammonia to investigate the effect. A more robust and automatic sensor would be needed, and this study has gained a lot of experience with different sensor technology as a result. The results overall are promising for being able to predict longevity of crossbred sows in the future, and for improving existing models in addition to the use of sensor technology. As a result, we expect increased genetic progress for longevity in crossbred sows. This should ideally lead to reducing the cost of replacement gilts for farmers due to less culling of sows due the problems with health and reproduction. Increased sales of semen internationally, due to an improved genetic product should benefit Norwegian pork producers. Healthier pigs, that are more robust will lead to better welfare and more sustainable production.

The project investigated ways to improve sow longevity. Longevity in crossbred sows is one of the most important traits for a piglet producer, and the Norwegian-Dutch TN70 sow is very popular both nationally and internationally for its mothering ability and good longevity. Thus, an improvement in longevity has the potential to increase national and international sales of pigs and semen. A data collection pipeline in crossbred herds was established, and a data collection pipeline for sensor data that allowed collection and storage was developed. This allowed for efficient and reliable data collection, which will impact the use of sensor data in the future. Experience and knowledge on the use of sensor data and machine learning methods furthered our understanding of longevity, related traits and sensor data. The knowledge gained will be a valuable contribution for further improving longevity in the future. Improving longevity will benefit pork producers due to reduced cost of replacement gilts and will benefit the sows through improved health and welfare. This also contributes to a more sustainable production, which benefits society. For NMBU, the project allowed researchers at the faculty of Science and Technology to further their understanding of pork production and develop models to work with a new type of data.

The TN70 sow is a cross between Landrace and Large White and is one of the most important animals in the pig production chain, as she produces the largest number of piglets and is very popular due to her excellent mothering ability. We aim to improve another trait in the TN70 sow, longevity, its ability to produce a high number of litters without reduced health and animal welfare. It is vital for the development of Norwegian swine genetics that we work even better towards improved longevity in the future. The trait is difficult to evaluate in purebred nucleus animals, as this would lead to a reduced turnover of animals, and thus a reduction in genetic gain. Solving this challenge within the current framework is tough. Several approaches to solve this problem have been attempted in the past with limited success. Part of the challenge is the differences that occur due to some animals being purebreds and some animals being crossbred. This creates statistical challenges due to differences in interactions between genes depending on whether they are from purebreds or crossbreds, and different interactions between genotypes and the environment. We will develop new methods for evaluating longevity through big data approaches such as machine learning and the inclusion of sensor technology to evaluate conditions at herd level. Building knowledge in the field of sensor technology and big data approaches is crucial for the future. The project findings will lead to increased genetic progress in longevity in the crossbred sows. This has the potential to increase international sales of pigs and semen with Norwegian genetics. For farmers, this will mean less culling of sows, and a reduction of replacement gilts, which is one of the largest costs. This will lead to improved economy for Norwegian pork producers, more sustainable food production, and more robust and healthier animals leading to better animal welfare. Less expensive food and improved food quality will benefit society.