Genetic analysis of traits derived from automated tracking and behaviour detection in pigs

Pigs contribute a major part to the animal protein production. As these animals are kept in groups during production it is essential to take their social interactions into consideration when making decisions on breeding. However, until now the inclusion of traits with regard to behaviour in the breeding goal is limited, mainly because behaviour is difficult to observe at the scale needed for breeding value estimations.

The use of artificial intelligence for animal tracking and behaviour
recognition based on image analysis is currently being developed in collaboration with the Eindhoven University of Technology and Clémence Orsini, another PhD candidate at the Adaptation Physiology group. This opens up opportunities to automatically score behaviour, allowing for large scale phenotyping of behaviour traits.

This thesis aims to look at traits derived from automated animal tracking and behaviour detection in order to understand their potential for genetic selection. Traits in question for this thesis are tail biting, social network metrics and activity. Furthermore, this project aims to find a way to incorporate social network parameters into models for breeding value estimation.

This project is part of the IMAGEN project that aims to create 1) algorithms to automatically track and detect animals, 2) analyse the detected behaviours and tracks from a behavioural science perspective and 3) look at the genetics of the detected behaviours and the tracks. The pigs that will be studied come from the breeding company Topigs Norsvin.