Epigenetic aging in a self-fertilizing vertebrate, the mangrove rivulus Kryptolebias marmoratus

Abstract

Background: Since Horvath (2013) and Hannum (2013), DNA methylation has emerged as a reliable biomarker for estimating age. Predictive models, called epigenetic clocks, are being built across the tree of life, including humans, other mammals and vertebrates and, recently, invertebrates. However, it always involves cross-fertilizing organisms, thus presenting a high level of genetic diversity. To decipher the exact role of epigenetics in adaptation and evolution, distinguishing true epigenetic variation from changes that reflect genetic variation is essential. Results: Here, we used the mangrove rivulus, the only self-fertilizing vertebrate, to study epigenetic aging in the brain and construct an epigenetic clock. We generated a reduced representation bisulfite sequencing dataset of 90 brain tissues from individuals aged 60-1100 days to construct a highly accurate age predictor using 40 CpG sites (R squared > 0.96, mean absolute error of 28.7 days). We associated the CpG sites with their respective closest genes and explored their biological functions. Genes relevant to aging include lamin-A (responsible for several age-related processes), aryl hydrocarbon receptor (whose absence causes premature aging) and several genes associated with Alzheimer\'s disease identified in human studies. We also propose several methods to improve classic epigenetic clock regression. Conclusions: This study is the first construction of an epigenetic clock in a self-fertilizing species, demonstrating that DNA methylation patterns undergo consistent changes across the lifespan of isogenic individuals. These findings emphasize putative changes in DNA methylation related to functional changes across aging in the brain.