IMAGEN – Behavioural interactions in laying hens

As regulations for animal production and welfare change, so do the methods needed for genetic analysis. With changes in the housing systems of the laying hen industry, welfare issues arose due to potential damaging behaviours (e.g., severe feather pecking or piling), these behaviours can increase the ecological footprint of egg industry. Therefore, an interest for breeding out these behaviours exist.For the expression of social behaviour, an interaction between individuals must occur. The genetic effects of these interactions are known as indirect genetics effects (IGEs). Until now, analyses considering IGEs have focussed on populations consisting of many small groups. However, in laying hen populations housed in cage-free facilities with larger groups, the identification of the individuals involved in an interaction is needed to estimate IGEs. Recent advances in sensing technologies, artificial intelligence and genomics provide radical new opportunities to address this issue. As part of the NWO-TTW Perspective programme IMAGEN, this project focuses on assessing the feasibility of genetic analysis of social behaviours accounting for IGEs, using automated phenotyping in laying hen populations based on computer vision. Genetic components for feather score of populations in cage-free housing systems will be estimated (estimated genetic components from traditionally housed populations will serve as reference). Automated individual trackability, and identification of behaviours will be assessed using a video set up, and from the obtained automated phenotypes, genetic components for social behaviours considering IGEs will be estimated. RStudio will be used for data management, ASRmel for genetic components estimation and MixBlup for EBV estimation. Developing continuous phenotyping methodologies using artificial intelligence will benefit animal welfare and will also allow a more accurate recording of behaviour. These phenotypes can be used for behavioural breeding, which can be translated into a lower ecological footprint as behaviourrelated mortality would be reduced.