Reference-free reduced representation bisulfite sequencing uses enzymatic digestion for reducing genome complexity and allows detection of markers to study DNA methylation of a high number of individuals in natural populations of non-model organisms. Current methods like epiGBS enquire the use of a higher number of methylated DNA oligos with a significant cost (especially for small labs and first pilot studies). In this paper, we present a modification of this epiGBS protocol that requires the use of only one hemimethylated P2 (common) adapter, which is combined with unmethylated barcoded adapters. The unmethylated cytosines of one chain of the barcoded adapter are replaced by methylated cytosines using nick translation with methylated cytosines in dNTP solution. The basic version of our technique uses only one restriction enzyme, and as a result, genomic fragments are integrated into two orientations with respect to the adapter sequences. Comparing the sequences of two chain orientations makes it possible to reconstruct the original sequence before bisulfite treatment with the help of standard software and newly developed software written in C and described here. We provide a proof of concept via data obtained from almond (Prunus dulcis). Example data and a detailed description of the complete software pipeline starting from the raw reads up until the final differentially methylated cytosines are given in Supplementary Material making this technique accessible to non-expert computer users. The adapter design showed in this paper should allow the use of a two restriction enzyme approach with minor changes in software parameters.

No abstract