Cytosine methylome data is commonly generated through next-generation sequencing, with analyses averaging methylation states of individual reads. We propose an alternative method of analysing single-read methylome data. Using this method, we identify patterns relating to the mechanism of two plant non-CG-methylating enzymes, CMT2 and DRM2. CMT2-methylated regions show higher stochasticity, while DRM2-methylated regions have higher variation among cells. Based on these patterns, we develop a classifier that predicts enzyme activity in different species and tissues. To facilitate further single-read analyses, we develop a genome browser, SRBrowse, optimised for visualising and analysing sequencing data at single-read resolution.
Keywords: DNA methylation; Epigenetic variation; Genome browser; NGS analyses.