Characterizing short-term evolution of DNA methylation in A. thaliana using next-generation sequencing

Abstract

DNA sequence mutations are the principal source of natural variation. Over the last few decades, however, an increasing number of studies have suggested that also epigenetic components can be at the basis of differences in phenotypic traits. These epigenetic marks allow a flexible modulation of gene activity without changes in the DNA sequence. One of the most prominent epigenetic modifications is DNA methylation, which consists of cytosines that carry an additional methyl group. Such chemical marks can be inherited across cell divisions and generations, and there are many durable methylation differences between individuals, so-called epimutations. These can originate from mainly three different sources: most epimutations are coupled to genetic mutations, yet they can also arise spontaneously, or they can be induced by environmental stimuli. The latter case enables rapid adaptation to changing environments, which in the short term is usually not possible via genetic mutations. A current debate revolves around the question whether adaptive environmentally induced epimutations can be heritable, which would contradict the random mutagenesis assumption of Darwinian evolutionary theory.

However, the experimental setup of most studies that have examined epigenetic variation did not allow the clear separation of different sources of variable methylation. These studies typically did not inspect genome-wide genetic variation, or did not monitor environmentally induced changes for more than one or two generations. Thus it has remained largely unresolved how frequently methylation differences arise spontaneously on the whole-genome level, and how strongly and durably environmental conditions impact the methylation landscape.

This work addresses these questions in the model plant Arabidopsis thaliana. I present whole-genome DNA methylation analyses at base-pair resolution of two different populations, originating from unique experimental settings that largely eliminate specific sources of epimutations. Investigation of genetically quasi identical lines propagated for thirty generations in uniform greenhouse conditions – thus largely without genetic and environmental influences – revealed that spontaneously occurring epimutations emerged frequently, but seemed to be largely short-lived. Plants with minimal genetic divergence that had grown in diverse natural sites over a previously uncharted time period of over one hundred years exhibited a methylation pattern that was largely stable on the whole-genome level and that was in many aspects intriguingly similar to that of the greenhouse-grown lines. Thus, environmentally induced epimutations seem to be only minor contributors to heritable methylation differences, which challenges published claims of broad-scale inheritance of adaptive epigenetic variation.

This thesis also provides technical and methodological advances of next-generation sequencing (NGS) data analysis. To gauge the genome-wide genetic influence on epimutations, this work provides an iterative workflow that maximizes the detection of a wide range of DNA sequence variants using short NGS reads by integrating several different genetic variation detection approaches. Finally, while previous epigenetic studies in plants, due to rather simplistic statistical testing, largely revealed a biased picture of differential methylation in the genome, this work introduces a comprehensive DNA methylation pipeline for NGS data that includes a novel approach to obtain more sensitive and more unbiased calls of differentially methylated regions.

Together, this work presents advanced computational methods to profile genome-wide genetic and methylation variation, and inspects the rate and spectrum of naturally occurring methylation changes, thus contributing to elucidating the role of epimutations in evolution.