Abstract
Epigenetic clocks are machine learning models that predict an organism's chronological age (the time elapsed since birth) or biological age (a proxy for physiological integrity) based on methylation levels from multiple genomic sites. To date, all epigenetic clocks rely exclusively on C5-methylcytosine (5 mC), the predominant DNA methylation mark in vertebrates. However, not all species possess detectable 5 mC levels. Here, we used N6-methyladenine (6 mA), a less-characterized DNA modification type, to develop a series of epigenetic clocks in the buff-tailed bumblebee (Bombus terrestris). Using long-read Nanopore sequencing, we generated genome-wide, base-resolution profiles of 6 mA and 5 mC in males of different ages (n = 15), and developed multiple epigenetic clocks based on distinct features of the aging DNA methylome. All clocks showed strong correlations between predicted epigenetic and chronological age. Moreover, they also detected pharmacologically induced lifespan extension, reflected by a reduction in predicted epigenetic age relative to chronological age, indicating that these clocks capture biological aging. These findings demonstrate that 6 mA can be used to build accurate epigenetic clocks and establish 6 mA as a promising biomarker of aging in animals.