Background/Question/Methods
The role of epigenetic variation in ecology and evolution depends on its contribution to diversity, trait variation among populations, and fitness in local environments. Average trait differences have been associated with epigenetic differences, but current methods for investigating epigenotype-phenotype correlations are limited. Classic approaches for understanding genotype-phenotype correlations partition trait variation and determine causes of variation at different levels (e.g., population and habitat). We suggest extending this approach to compare epigenetic, and phenotypic structure among populations and habitats. Most analyses of epigenetic structure have used measures of DNA methylation (e.g., restriction enzymes) which represent small portions of the epigenome and do not test correlations with phenotypic structure. Enrichment-based sequencing of DNA methylation covers a large portion of the genome and costs less than other genome-wide methods. It has been widely used in both ecological and disease contexts. We introduce a method for quantifying epigenetic structure and its association with phenotypic structure using enrichment-based sequencing data. As proof of concept, we examine the role of epigenetic variation in explaining trait divergence of a clonal snail between lakes and rivers in its North American invasive range. First, we test whether populations with divergent traits have higher structure than populations with similar traits. Next, we show how Mantel and partial Mantel tests can be used on pairwise epigenetic and phenotypic distances to identify regions of the methylome most correlated with phenotypic divergence.
Results/Conclusions
Our results suggest that DNA methylation is significantly structured by population and habitat. Additionally, the structure observed between habitats with distinct phenotypes exceeds the structure among populations. This observation is consistent with a role of epigenetic variation in phenotypic divergence. We also identified regions of the methylome most differentiated by habitat (i.e., epigenetic outliers) and most correlated with phenotypic divergence. Altogether, the framework we presented provides the ability to partition sources of variation in DNA methylation and determine its causes. This was previously not possible with popular enrichment-based sequencing methods such as Methylated DNA Immunoprecipitation sequencing and Methyl Binding Domain capture sequencing. When combined with trait data, our method can be used to directly evaluate correlations between divergence in methylation and phenotypic variation. Because enrichment-based sequencing is used widely in ecological epigenetics and oncology, this approach could help identify epigenetic underpinnings of trait variation in both wild populations and populations of cancer cells.