In addition to playing a role in genomic function, DNA methylation influences evolution by regulating transcription. Technological advances, such as High-Performance Liquid Chromatography (RP-HPLC), have allowed scientists to explore genomic regulatory changes that contribute to species diversity and phenotypic variability. Epigenetic modifications of notable interest include 5-methylcytosine (5mC) and CpG (GC), as they are related to neutral selection on the cellular level. To understand how these regulatory changes evolve in a phylogenetic context, quantitative traits were analyzed phylogenetically by mapping them to separate mitochondrial phylogenies inferred de novo across 28 reptile species, 26 mammal species, and 42 fish species. Previous studies in vertebrates concluded that there was no significant correlation between DNA methylation and environmental stimuli, but these studies did not correct for the non-independence of evolutionarily related species and thus violated a fundamental statistical assumption. To model these traits, traits were corrected for phylogeny and phylogenetic comparative analyses were run in RStudio®. First, the extent of the phylogenetic non-independence problem was examined by estimating measures of phylogenetic signal for each quantitative trait. Then regressions were repeated from a previous study, following phylogenetic correction, and inferred correlation between our two epigenetic modifications. Finally, a series of evolutionary models were fit to the phylogeny to examine the evolution of these traits across the phylogenies and selected the best fit model using an AICc model selection procedure. Phylogenetic signal was found in 5mC for both mammal and fish species. In reptile species, phylogenetic signal was found in GC but not 5mC, and that phylogenetic correction did not affect results, likely owing to the relatively small number of tips and the lack of phylogenetic signal in one of the traits. The evolution of these traits is best approximated by an Ornstein-Uhlenbeck model, suggesting that local optima exist for these quantitative characters and predicting a loss of phylogenetic signal (convergence or homoplasy). This study is important because the results can be used to understand the modifications to the genome influencing phenotypic diversity.
