Interpretive Summary: Genomic selection in the era of digital phenotyping

By: Daniela Lourenco, Matias Bermann, Mary Kate Hollifield, Masum Billah, Ching-Yi Chen, Eric Psota, Justin Holl, Shogo Tsuruta, Ignacy Misztal (University of Georgia)

Promoting sustainable breeding programs requires several measures, including genomic selection and continuous data recording. Digital phenotyping can be used to track animal activity and behavior like feeding and walking time and distress continuously. Coupled with machine learning techniques, any feature of interest can be extracted and used as phenotypes in genomic prediction models. It can also help define novel phenotypes that are hard or expensive to measure by humans. For the already recorded traits, it may add extra precision or lower phenotyping costs. One example is lameness in pigs, where digital phenotyping allowed moving from a categorical scoring system to a continuous phenotypic scale, resulting in increased heritability and greater selection potential. Additionally, if an early digital phenotyping behavior is genetically correlated with a trait of interest measured later, selection decisions can be made earlier. One example is the strong, negative genetic correlation (-0.7) between distance traveled and average daily gain in pigs. Conversely, computer vision may add noise to the phenotype for some production or carcass traits, as correlations with the traditional records can be as low as 0.9. In this talk, we will review the benefits and opportunities of digital phenotyping, together with experiences in data analysis and quality control. Finally, we will discuss how to account for the inaccuracy of some phenotypes in genomic prediction models. Overall, digital phenotyping is a promising tool to increase the rates of genetic gain, promote sustainable genomic selection, and lower phenotyping costs.