Nanopore duplex sequencing reveals patterns of asymmetric states of 5hmC and 5mC in the medaka brain genome

The nucleotide modification 5-methylcytosine (5mC), has been extensively described in terms of its genomic distribution and function. However, the distribution and functional contributions of its oxidised state, 5-hydroxymethylcytosine (5hmC), and its interplay with 5mC remain to be further characterised. Moreover, the hemi-methylation and hemi-hydroxymethylation of CpG dinucleotides in the genome also remain largely unexplored. Here, we use Nanopore duplex sequencing, i.e, the sequencing of both strands of DNA from a single double stranded molecule, to profile the co-occurrence of 5hmC with 5mC in CpG dinucleotides of medaka (japanese rice paddy fish, Oryzias latipes) brain tissues. We characterised asymmetrical patterns of hemi-modified states of CpGs and show using duplex resolution that in the asymmetrical state 5hmC/5mC, 5hmC is preferentially found in the antisense strand of DNA in intron regions, in particular in splice sites, whereas the exons have an enrichment in 5hmC in the sense strand. Furthermore, we found that in the 5hmC/5mC state, 5hmC preferentially occurs along one strand of the DNA duplex in specific families of transposable elements. Altogether these results shed more light on understanding the function and distribution of 5hmC and its mutual role with 5mC in the genome.