Analysis of variation in epigenetic age prediction in Domestic dog samples (Canis lupus familiaris)

Nirali Desai1,2, Matteo Pellegrini2 1College of Science, Northeastern University, Boston, MA 25000D Terasaki Life Sciences Building, Molecular, Cell & Developmental Biology Department, University of California, Los Angeles, California Corresponding Author Matteo Pellegrini 5000D Terasaki Life Sciences Building, Molecular, Cell & Developmental Biology Department, University of California, Los Angeles, California Email: matteope@gmail.com

Epigenetic clocks, also known as DNA methylation clocks, are biomarkers of aging and consist of CpG sites whose DNA methylation levels change with age in humans and other vertebrates. Some studies suggest that these clocks generate predictions with high variance in repeated samples collected within narrow time spans. Therefore, the goal of the study was to evaluate various epigenetic clocks using machine learning approaches to measure the degree to which they impact the variance of age predictions. The study was conducted using buccal swabs from canids. We previously collected 200 samples of dogs of different ages, along with 16 technical replicates of the sample which will be used to assess the intraday variance of age predictions. Our goal is to optimize age prediction both based on minimizing age prediction error as well as the variance of replicate samples.