Improving bisulfite short-read mapping efficiency with hairpin-bisulfite data

DNA methylation is an important epigenetic mark relevant to normal development and disease genesis. A common approach to characterizing genome-wide DNA methylation is to use Next Generation Sequencing technology to sequence bisulfite treated DNA. The short sequence reads are mapped to the reference genome to determine the methylation status of Cs. However, despite intense effort, a much smaller proportion of the reads derived from bisulfite treated DNA (usually about 40-80%) can be mapped than regular short reads mapping (≥ 90%), and it is unclear what factors lead to this low mapping efficiency. To address this issue, we used the hairpin sequencing technology to determine sequences of both DNA double strands simultaneously. This enabled the recovery of the original non-bisulfite-converted sequences. Recovered reads had a higher unique mapping efficiency than the bisulfite reads, and one reason could be that mapping efficiency relates to entropy, and entropy may be lost due to bisulfite treatment. Bismark was used to map bisulfite reads, and Bowtie2 was used to map recovered untreated reads.