Regulome

Several DNA methylation analysis methods have been developed to assess the status of numerous CpG islands or the entire genome. The most widely used, bisulfite conversion of unmethylated cytosines to uracil, presents major drawbacks: DNA degradation, complex library preparation, conversion bias, and an inability to directly distinguish 5-methylcytosine (5mC) from 5-hydroxymethylcytosine (5hmC).

Messenger RNA also harbors modifications (epitranscriptome), the most prevalent in eukaryotes being N6-methyladenosine (m6A), which regulates RNA splicing, translation, and stability. Traditionally quantified via immunoprecipitation (m6A-seq), this approach suffers from limited specificity, antibody bias, and low yields (Deng X. et al.  2023).

To overcome these hurdles, long-read sequencing technologies (PacBio HiFi and Oxford Nanopore Technologies – ONT) enable the direct analysis of native DNA and RNA without amplification or sample-damaging chemical treatments.

Modifications (5mC, 5hmC, 6mA on DNA; m6A, pseudouridine on RNA) are detected directly as the molecule translocates through a nanopore or is processed by a polymerase. Physical signal variations; interpulse duration (IPD) for PacBio (Fig. A) or ionic current disruptions for ONT (Fig. B) ; deviate from canonical distributions. (figure from van Dijk E.L. et al. 2023).

Bioinformatic analysis has achieved a major breakthrough. Once reliant on comparing raw signals (Tombo, Nanopolish), modification calling now leverages deep learning models integrated directly into real-time base-calling (Dorado and Remora from ONT, Primrose from PacBio). Improved single-read accuracy (notably with ONT R10.4.1 chemistry and the RNA004 direct RNA kit) enables simultaneous, multimodal detection of diverse epigenetic marks at single-base resolution. Furthermore, mapping these modifications across very long native reads facilitates haplotype phasing of the epigenome and epitranscriptome within complex or repetitive genomic regions.

Update: Aug 2026