Fig. 1: Making use of population divergence to analyze the source of epigenetic variation.

Genealogical divergence (e.g. between individuals, families, populations, species) is typically reflected by the amount of difference in (heritable) genetic variation (Δ genetic). Contrasting heterogeneity in genetic and environmental divergence helps in assessing the degree of autonomous epigenetic variation (Box 2). a Three population samples with increasing genealogical distance are depicted (A, B), C). B1 (blue) and B2 (yellow) have the same genetic background, but are subject to different environmental conditions. b Associations between genome-wide genetic and epigenetic variation can be assessed using the correlation of distance metrics for genetic (Δ genetic) and epigenetic (Δ epigenetic) variation among genetic clusters, or through multivariate statistical approaches such as principle component analyses (PCA, not shown). For genetically controlled epigenetic variation, we expect to see a strong correlation between genome-wide metrics of genetic and epigenetic distance regardless of the environment. High correlation of genetic and epigenetic distance matrices between pairs of populations, and not the environment, may be a first indication of genetic control. c) While single-value comparisons per population provide a mean estimate across populations, window-based measures allow capturing variation along the genome. Capitalizing on the heterogeneity of genetic and epigenetic variation along the genome further allows to pinpoint regions of the genome under genetic and/or environmental control. The upper panel shows variation in genetic and epigenetic distance for the population comparison A | B2 at a regional scale along the genome. A point can refer to any genomic window in units of base pairs (or better units of centimorgans) here referred to as a locus. The lower panel shows an example of genetically controlled (left) and environmentally associated (right) epigenetic locus. Δ genetic is shown in blue, Δ epigenetic in red. Box 4 for corresponding statistical approaches.