Long-Read Sequencing of RNA Methylation — Direct Detection of m⁶A, m⁵C & Ψ at Single-Molecule Resolution by Nanopore Direct RNA Sequencing

Long-Read Sequencing of RNA Methylation — Direct Detection of m⁶A, m⁵C & Ψ at Single-Molecule Resolution by Nanopore Direct RNA Sequencing

Long-read sequencing of RNA methylation — direct m6A detection by Nanopore direct RNA sequencing for epitranscriptome analysis

CD Genomics provides Nanopore direct RNA sequencing for RNA methylation analysis — detecting m⁶A, m⁵C, pseudouridine (Ψ), and inosine directly from native RNA without reverse transcription, PCR amplification, or antibody enrichment. Single-molecule resolution enables isoform-resolved modification analysis, modification stoichiometry, and simultaneous poly(A) tail profiling from the same sequencing run.

RNA methylation is the most abundant and functionally diverse class of post-transcriptional modification in the epitranscriptome. Over 170 distinct RNA modifications have been characterized across all RNA biotypes — messenger RNA (mRNA), transfer RNA (tRNA), ribosomal RNA (rRNA), and non-coding RNA — where they regulate splicing, translation, RNA stability, and cellular stress responses. Among these, N⁶-methyladenosine (m⁶A) is the most prevalent internal mRNA modification, affecting over 60% of human transcripts and modulating every stage of RNA metabolism from nuclear export to decay. 5-methylcytosine (m⁵C), pseudouridine (Ψ), and inosine (I) expand the functional repertoire, yet the vast majority of modification sites remain incompletely mapped across transcriptomes, isoforms, and cell types — because the dominant detection methods have, until recently, relied on antibody enrichment or chemical conversion that obscures single-molecule resolution and isoform-level information.

Long-read direct RNA sequencing (dRNA-seq) on the Oxford Nanopore Technologies (ONT) platform eliminates these constraints entirely. By sequencing native RNA molecules without reverse transcription or PCR amplification, dRNA-seq preserves every endogenous modification in its original context — the ionic current signature generated as each RNA molecule passes through a nanopore carries a characteristic shift when a modified base is present, enabling simultaneous detection of m⁶A, m⁵C, Ψ, and inosine from the same sequencing run. Every read delivers full-length transcript sequence, isoform structure, modification status at single-nucleotide resolution, and poly(A) tail length from the same native RNA molecule. This is not a surrogate for antibody-based epitranscriptomics — it is a fundamentally different information layer that captures the complete modification landscape of each transcript isoform individually, rather than population-averaged enrichment signals.

Why Long-Read Direct RNA Sequencing for RNA Methylation — Service Highlights

Why Long-Read Direct RNA Sequencing for RNA Methylation — and Why the Epitranscriptome Needs Single-Molecule Resolution

RNA methylation research has been transformed over the past decade by the development of transcriptome-wide m⁶A mapping methods — MeRIP-seq (m⁶A antibody enrichment followed by short-read sequencing), m⁶A-CLIP, and m⁶A-EXO-seq — that have established m⁶A as a widespread, dynamically regulated RNA modification with essential roles in development, stress response, and disease. These methods share a fundamental limitation: they report population-averaged modification signals across thousands of cells and transcript molecules. An m⁶A peak called by MeRIP-seq tells you that a region of a transcript is enriched for m⁶A in the bulk cell population, but it cannot tell you what fraction of the transcript molecules carry the modification at that site, how m⁶A at different sites co-occurs on the same molecule, or how modification status differs across transcript isoforms originating from the same gene.

Long-read direct RNA sequencing (dRNA-seq) addresses every one of these limitations. By sequencing individual native RNA molecules end-to-end through a nanopore, dRNA-seq measures the modification status of every nucleotide on every captured molecule, producing a single-molecule modification map of the transcriptome. This single-molecule resolution unlocks four categories of biological information that no population-averaged method can provide: (1) modification stoichiometry — the fraction of transcripts carrying a modification at each site, measured directly from per-read modification calls rather than inferred from enrichment ratios; (2) modification co-occurrence — whether m⁶A sites on the same transcript are modified independently or coordinately, and how modification status at one site relates to another; (3) isoform-specific modification — whether alternative transcript isoforms from the same gene carry systematically different modification patterns, and how these patterns relate to isoform-specific functions; and (4) modification-transcript feature relationships — how m⁶A status relates to poly(A) tail length, alternative splicing, and RNA abundance on a per-molecule basis, all from the same native RNA reads. The Nanopore Direct RNA Sequencing service page provides detailed information on the core technology platform.

Our service is designed to make this single-molecule epitranscriptomics capability accessible to researchers studying RNA methylation across any biological system — from focused analysis of specific modification types in targeted transcript sets to comprehensive whole-transcriptome m⁶A, m⁵C, and Ψ profiling at isoform resolution.

Direct Long-Read RNA Modification Detection Uses Nanopore Ionic Current Signatures Rather Than Antibody Enrichment or Chemical Conversion

Direct long-read RNA modification detection utilizes the Oxford Nanopore direct RNA sequencing (dRNA-seq) platform, which sequences native RNA molecules without reverse transcription, PCR amplification, or chemical labeling. The detection principle is based on physical signal measurement rather than biochemical conversion or antibody recognition.

Nanopore direct RNA sequencing with current-based modification detection: During dRNA-seq, a native RNA molecule is captured by a motor protein and translocated through a nanopore embedded in an electrically resistant membrane. As each nucleotide passes through the pore's constriction, it modulates the ionic current flowing across the membrane in a sequence- and modification-dependent manner. The current signal — measured at thousands of measurements per second — is characteristic not only of the canonical base (A, C, G, U) but also of any chemical modification present on the base. Modified bases such as m⁶A, m⁵C, Ψ, and inosine produce distinct current signatures that differ from their unmodified counterparts. These signatures are recognized during signal processing and basecalling through computational models trained on known modification states. The ONT PromethION flow cells (RNA004 chemistry) generate tens of millions of direct RNA reads per run, each ranging from 500 bp to over 10 kb depending on RNA integrity, providing comprehensive coverage of full-length mRNA transcripts.

Modification detection tools and models currently deployed in our pipeline: RNA modification calling from dRNA-seq data employs a tiered approach. At the basecalling level, ONT's Dorado v5+ modified-base models directly identify m⁶A (in DRACH motif contexts), pseudouridine, and inosine during the basecalling step, producing modified-base probability scores per nucleotide. For comprehensive transcriptome-wide m⁶A calling beyond DRACH motifs, we deploy m6Anet (a multiple-instance learning deep neural network) which provides per-read m⁶A probability scores at single-nucleotide resolution. For simultaneous m⁶A and m⁵C detection, we offer CHEUI (a two-stage neural network trained on endogenous labeling data), which predicts both modification types from the same sequencing run. All modification calls are delivered in MM/ML SAM tag format for compatibility with downstream tools including Minimod for vendor-agnostic modification analysis and visualization. Our Oxford Nanopore Sequencing Data Analysis page describes the full bioinformatics pipeline in detail.

What distinguishes direct RNA modification detection from alternative approaches: Unlike antibody-based methods (MeRIP-seq, m⁶A-CLIP) that report enriched regions in pooled populations of RNA fragments, dRNA-seq reports the modification status of every nucleotide on every individual RNA molecule captured. Unlike chemical conversion methods (bisulfite sequencing for m⁵C, CMC- labeling for Ψ) that require harsh reaction conditions and destroy RNA integrity, dRNA-seq reads native RNA under mild buffer conditions. Unlike short-read sequencing of fragmented RNA, dRNA-seq reads full-length transcripts end-to-end, preserving the linkage between modification sites on the same molecule and enabling isoform-resolved analysis. These differences are not incremental — they represent a categorical shift in the resolution and completeness of epitranscriptome analysis.

Nanopore dRNA-seq Provides Single-Molecule, Multi-Modification Detection with Integrated Poly(A) Tail and Isoform Analysis

Scientific Advantages

  • Multi-modality single-molecule epitranscriptomics

Every dRNA-seq dataset simultaneously captures m⁶A, m⁵C, Ψ, and inosine modification status from the same native RNA reads — no separate antibody enrichments, chemical treatments, or parallel sequencing runs are required. Per-read modification calls provide direct measurement of modification stoichiometry, co-occurrence, and isoform specificity.

  • Full-length isoform-resolved modification analysis

Long reads span entire transcripts end-to-end, linking 5′ UTR, CDS, and 3′ UTR modifications to specific transcript isoforms. Alternative promoters, splice variants, and APA isoforms that carry systematically different modification patterns are resolved individually — information that is structurally invisible when modifications are mapped to fragmented short reads aligned against a collapsed gene model.

  • Integrated RNA feature profiling from a single experiment

Every dRNA-seq read simultaneously reports full-length transcript sequence, isoform structure, base modification status, and poly(A) tail length. Correlational analyses — m⁶A vs poly(A) tail length, modification vs isoform usage, modification vs RNA abundance — are performed on the same native RNA molecules rather than integrated across separate experiments with different protocols and batch effects.

Business & Project Advantages

  • Single-assay epitranscriptome profiling — replace four parallel experiments

A single dRNA-seq experiment replaces separate RNA-seq (expression), isoform-seq (alternative splicing), modification-seq (m⁶A/MeRIP), and poly(A)-tail assays. This consolidation reduces both per-project cost and required RNA input, making comprehensive epitranscriptome analysis feasible for projects with limited sample material.

  • Integrated RNA modification service modules covering the full epitranscriptome

Our RNA methylation service integrates with complementary long-read RNA analysis services — Nano-tRNA Sequencing for specialized tRNA modification profiling, TAIL Iso-Sequencing for comprehensive poly(A) tail analysis with modification detection, and Full-Length Transcript Sequencing (Iso-Seq) for PacBio-based isoform characterization — enabling multi-layered epitranscriptome analysis from a single service provider.

  • End-to-end bioinformatics with validated modification callers

Our computational team deploys platform-optimized RNA modification detection pipelines: Dorado v5+ modified-base models for direct m⁶A/Ψ/inosine calling, m6Anet for comprehensive transcriptome-wide m⁶A detection, CHEUI for simultaneous m⁶A and m⁵C analysis, and Minimod for MM/ML-tag-based downstream analysis. Deliverables include per-site modification frequency tables, per-read modification status, differential modification analysis, isoform-level quantification, and integrated reports.

ONT PromethION with RNA004 Chemistry Is the Only Platform for Direct RNA Modification Detection at Transcriptome Scale

Long-read RNA methylation detection is performed exclusively on the Oxford Nanopore PromethION platform using direct RNA sequencing chemistry. Unlike DNA methylation analysis, which benefits from dual PacBio/ONT platform options, RNA methylation detection requires native RNA sequencing without reverse transcription — a capability that PacBio SMRT sequencing does not currently support. Our service is optimized around ONT PromethION with RNA004 chemistry as the primary detection platform.

Feature ONT PromethION (RNA004) Alternative Approaches
Detection principle Ionic current shift through nanopore — direct RNA sequencing without RT/PCR MeRIP-seq (antibody enrichment + short-read seq); bisulfite conversion (m⁵C); CMC labeling (Ψ)
Modifications detectable (simultaneously) m⁶A (DRACH + non-DRACH via m6Anet), m⁵C, Ψ, inosine, 2′-O-methylation Single modification per assay — separate experiments for each modification type
Resolution Single-nucleotide, single-molecule — per-read modification status at every position Population-averaged enrichment peaks (MeRIP-seq); single-nucleotide but bulk (bisulfite)
Isoform-level information ✓ Full-length native transcripts — isoform-resolved modification calls ✘ Fragmented RNA — isoforms collapsed; modification signal averaged across isoforms
Poly(A) tail simultaneously measured ✓ Per-read poly(A) tail length from the same native RNA molecule ✘ Requires separate poly(A)-tail assay
Read length (typical) 500 bp–10+ kb (native RNA, dependent on RNA integrity) 50–300 bp (short-read fragmented RNA)
Throughput per flow cell 5–15 million reads (RNA004 chemistry, PromethION) N/A — not directly comparable across orthogonal methods
RNA input requirement ≥ 1 µg polyA+ RNA or total RNA (optimized RNA004 chemistry supports 500 ng–1 µg) 5–20 µg total RNA (MeRIP-seq); 1–10 µg (bisulfite-seq)
Key bioinformatics tools Dorado v5+ mod-base, m6Anet, CHEUI, Nanopolish, Minimod MACS2 peak calling; exomePeak; bisulfite mapping tools

Our Service Modules Combine dRNA-seq with tRNA Modification, Poly(A) Tail, and Isoform-Level Analysis

Our long-read RNA methylation analysis extends beyond standard dRNA-seq modification calling. We offer a suite of integrated service modules that combine RNA modification detection with complementary epitranscriptomic measurements, enabling comprehensive multi-layer epitranscriptome characterization from a single service provider. Each module is available as a standalone service or as an integrated component of a broader RNA analysis project.

Nanopore Direct RNA Sequencing — Core RNA Modification Detection Service

Core service: Whole-transcriptome direct RNA sequencing on ONT PromethION (RNA004 chemistry) with multi-modification detection. Native polyA+ RNA is sequenced directly — no reverse transcription, no PCR amplification, no antibody enrichment — and base modifications (m⁶A, m⁵C, Ψ, inosine) are called from the ionic current signal using Dorado v5+ modified-base models, complemented by m6Anet for comprehensive m⁶A detection and CHEUI for simultaneous m⁶A/m⁵C analysis. Per-read modification status, single-nucleotide modification probabilities, and isoform-resolved modification profiles are delivered for every transcript detected. This core module is suitable for projects ranging from whole-transcriptome epitranscriptome discovery (baseline m⁶A landscape across tissues or conditions) to focused modification analysis of specific gene families or biological pathways. The Nanopore Direct RNA Sequencing service page provides detailed information on the core method, sample requirements, and standard deliverables.

Nano-tRNA Sequencing — Specialized tRNA Modification and Fragment Profiling

Nano-tRNA Sequencing applies Nanopore direct RNA sequencing to the most densely modified class of cellular RNA. Transfer RNAs carry the highest density and diversity of post-transcriptional modifications of any RNA biotype, with 15–30 modified residues per molecule on average — including m⁵C, pseudouridine, dihydrouridine, 1-methyladenosine (m¹A), 7-methylguanosine (m⁷G), and dozens of base- and ribose-methylations that collectively regulate tRNA folding, stability, aminoacylation, and codon decoding. Nano-tRNA-seq preserves these modifications through direct native tRNA sequencing and reports modification status at single-nucleotide resolution on individual tRNA molecules, enabling tRNA-isotype-resolved modification analysis, detection of hypo- and hyper-modified tRNA populations, and characterization of tRNA-derived fragments (tRFs) with their modification status intact.

TAIL Iso-Sequencing — Poly(A) Tail Length, 3′ UTR Modification, and APA Analysis

TAIL Iso-Sequencing combines direct RNA sequencing with specialized library preparation to capture full-length mRNA molecules with intact poly(A) tails, enabling simultaneous measurement of poly(A) tail length distribution, m⁶A modification status along the transcript body and 3′ UTR, alternative polyadenylation (APA) site usage, and transcript isoform structure from the same native RNA molecules. The per-read poly(A) tail length measurement enables direct correlation between m⁶A modification status and poly(A) tail length at single-molecule resolution — a relationship that published studies (Kim et al. 2025, Cell Genomics; see Case Study below) have shown to be more complex and context-dependent than previously understood. This integrated module replaces separate poly(A)-tail assays and m⁶A-seq experiments with a single sequencing run, preserving the molecular linkage between tail length and modification status on every captured transcript.

Full-Length Transcript Sequencing (Iso-Seq) — Isoform-Level Modification Context

Full-Length Transcript Sequencing (Iso-Seq) on the PacBio Revio platform provides complementary high-accuracy (Q30+) full-length transcript sequences that serve as reference isoform annotations for RNA modification analysis. While PacBio Iso-Seq does not directly detect RNA modifications (it requires RT and PCR), the comprehensive isoform catalog it generates — including novel isoforms not represented in reference annotations — enables more accurate assignment of m⁶A calls to specific transcript isoforms from dRNA-seq data. This service module is recommended for projects in non-model organisms or poorly annotated genomes where the isoform reference is incomplete, and for cross-platform validation studies requiring independent isoform-level confirmation of modification-associated alternative splicing events.

Long-Read RNA Methylation Sequencing Addresses Cancer, Developmental, Neuroepigenetic, Plant, and Viral Epitranscriptome Research

Beyond well-established areas in cancer and developmental biology, long-read RNA methylation sequencing is increasingly applied in aging and longevity research, where single-molecule resolution enables detection of age-related m⁶A drift across transcript isoforms and correlation with poly(A) tail shortening that population-averaged methods cannot resolve. In chemical biology and drug development, dRNA-seq provides direct readout of target engagement for RNA-modifying enzyme inhibitors — measuring m⁶A reduction at individual sites following METTL3/METTL14 inhibitor treatment — supporting epitranscriptome-targeted therapeutic development with single-molecule pharmacodynamic data that antibody-based methods cannot provide at the required resolution.

Cancer Epitranscriptomics

  • Single-molecule m⁶A profiling in tumor transcriptomes to identify modification changes in oncogene and tumor suppressor transcripts, including isoform-specific m⁶A differences — such as METTL3-dependent modification shifts that regulate transcript stability and translation in drug-resistant cancer cell populations as demonstrated by direct RNA-seq studies in chemotherapy-resistant breast cancer cells
  • Integrated analysis of m⁶A modification, poly(A) tail length, and alternative splicing from the same native RNA reads, enabling multi-layered characterization of epitranscriptomic dysregulation in cancer without the confounding effects of cross-experimental batch variation

Developmental and Stem Cell Biology

  • Isoform-resolved m⁶A dynamics during stem cell differentiation, embryonic development, and cellular reprogramming — where population-averaged MeRIP-seq signals obscure cell-type-specific modification patterns and isoform-level regulatory logic
  • Allele-specific m⁶A analysis in hybrid or F1 systems, where long reads spanning heterozygous RNA SNPs directly phase modification calls to parental alleles, revealing imprinting-like modification patterns in the transcriptome

Neuroepitranscriptomics and Brain Disorders

  • Whole-transcriptome m⁶A and Ψ profiling in post-mortem brain tissue and neuronal cell models to characterize modification dynamics at synaptic plasticity genes, imprinting-associated loci, and neurological disease risk genes
  • Single-molecule modification analysis of long neural transcripts (>5 kb) that carry critical regulatory m⁶A sites but are systematically underrepresented in short-read fragmented RNA sequencing due to length-dependent fragmentation bias

Plant Epitranscriptomics and Stress Response

  • Transcriptome-wide m⁶A and m⁵C mapping in plant species under abiotic stress (heat, drought, salinity) and during developmental transitions (flowering, seed germination), leveraging long native reads that span entire plant mRNAs including full-length UTRs where regulatory m⁶A sites are enriched
  • Simultaneous measurement of m⁶A modification, alternative polyadenylation, and poly(A) tail dynamics in stress-responsive transcripts, capturing the coordinated regulation of RNA modification and RNA processing in the same native molecules

Microbial and Viral Epitranscriptomics

  • Direct detection of m⁶A and other modifications in viral RNA genomes and transcripts — including SARS-CoV-2, HIV-1, and ZIKV — where long-read dRNA-seq provides the only method that can map modifications to specific viral RNA isoforms without the bias introduced by RT-PCR amplification of structured viral RNA
  • Bacterial and archaeal RNA modification profiling, including modification-based regulation of rRNA and tRNA function in antibiotic resistance mechanisms and stress-adaptive translational control

Our Pipelines Deliver Per-Site and Single-Molecule RNA Modification Calls with Differential and Isoform-Level Analysis

Analysis Feature Standard Package Advanced Package
Read preprocessing, basecalling (Dorado v5+ with RNA004 modified-base models), and QC
Per-site modification frequency (m⁶A, Ψ, inosine) with probability scores (MM/ML SAM tag format)
Per-read single-molecule modification status
Transcriptome-wide m⁶A detection via m6Anet (beyond DRACH motif)
Simultaneous m⁶A and m⁵C detection via CHEUI
Differential modification analysis between conditions (per-site and per-isoform)
Isoform-level modification quantification and isoform-specific modification comparison
Per-read poly(A) tail length estimation and m⁶A–poly(A) correlation analysis
Modification co-occurrence analysis (e.g., m⁶A + Ψ on same read)
Functional annotation and pathway enrichment of differentially modified transcripts
Custom downstream analysis and publication-ready figures

Sample Input Requirements for dRNA-seq — RNA Quality, Input Amount, and Coverage Targets for RNA Modification Detection

Category Requirement Notes
Sample type Total RNA or polyA⁺-enriched RNA (tissue, cells, or purified RNA) Native, unmodified RNA — no RT, no PCR, no chemical treatment
Minimum input (polyA⁺ RNA) ≥ 1 µg (standard); 500 ng (optimized, RNA004 chemistry) PolyA⁺ enrichment recommended for mRNA-focused studies; rRNA depletion available for total RNA
Minimum input (total RNA) 5–10 µg (standard); 2–5 µg (optimized protocol) Higher input compensates for lower mRNA fraction in total RNA; rRNA depletion recommended
RNA quality RIN ≥ 8; A260/280 ≥ 1.9; A260/230 ≥ 1.8; no visible degradation Assessed by Bioanalyzer or TapeStation; degraded RNA reduces full-length read yield and modification detection sensitivity
Coverage recommendation 5–15 million reads per sample (whole-transcriptome); 1–3 million (targeted transcript set) Higher coverage required for low-abundance transcript detection and rare modification analysis
Shipping conditions Overnight on dry ice (tissue/pellet); liquid nitrogen vapor (purified RNA in TRIzol or RNAstable) See our sample submission guidelines for detailed instructions

CD Genomics Provides dRNA-seq as a Core Platform with Validated Multi-Modification Bioinformatics Pipelines

Direct RNA sequencing is our core transcriptomics platform, not an add-on service.

CD Genomics has built its long-read transcriptomics service portfolio around ONT direct RNA sequencing from the ground up. Our RNA methylation analysis workflows are designed for native RNA — no RT, no PCR, no workarounds that compromise modification detection. When you work with us, your epitranscriptome project is executed by a team whose core technology platform matches the native RNA approach your project requires.

Multi-modification detection from a single sequencing run — no parallel experiments.

We detect m⁶A, m⁵C, Ψ, and inosine simultaneously from the same dRNA-seq dataset, using a tiered bioinformatics approach that combines Dorado v5+ modified-base models, m6Anet, CHEUI, and Minimod. You receive comprehensive transcriptome-wide modification profiles from a single sequencing run and a single RNA input — not from separate antibody enrichments, chemical conversions, or parallel library preparations that multiply cost and sample requirements.

Platform-appropriate bioinformatics for RNA modification data.

RNA modification calling from dRNA-seq requires fundamentally different algorithms than DNA methylation detection. Our bioinformatics team deploys pipelines specifically designed and validated for direct RNA modification detection — including m6Anet (multiple-instance learning for transcriptome-wide m⁶A), CHEUI (two-stage neural network for m⁶A + m⁵C from the same data), and Dorado modified-base models for direct basecalling-level modification identification. We do not apply DNA methylation pipelines to RNA data or use generic peak-calling approaches designed for antibody-enriched fragments.

Integrated multi-module epitranscriptome analysis from a single provider.

Beyond whole-transcriptome modification detection, our integrated service modules — Nano-tRNA Sequencing for tRNA modification profiling, TAIL Iso-Sequencing for simultaneous modification + poly(A) tail analysis, and Iso-Seq for complementary high-accuracy isoform reference annotation — enable comprehensive, multi-layered epitranscriptome characterization without coordinating across multiple service providers or reconciling data from incompatible platforms.

Validated through peer-reviewed publications.

Our direct RNA sequencing services have been used in published studies spanning cancer epitranscriptomics, plant stress RNA biology, and viral RNA modification analysis, with datasets generated on our platforms appearing in peer-reviewed journals.

Case Study: Single-Molecule m⁶A Detection by Direct RNA Sequencing in Leukemia Cells — Simultaneous Modification, Poly(A) Tail, and Isoform Analysis

Kim Y, Saville L, O'Neill K, Garant JM, Liu Y, Haile-Merhu S, Ghashghaei M, Hoang QA, Louwagie A, Park YP, Jones SJM, Vu LP. Nanopore direct RNA sequencing of human transcriptomes reveals the complexity of mRNA modifications and crosstalk between regulatory features. Cell Genomics. 2025;5(6):100872. (CC BY 4.0)

1. Background

N⁶-methyladenosine (m⁶A) is the most abundant internal mRNA modification in eukaryotes, yet the relationships between m⁶A modification, poly(A) tail length, mRNA abundance, and alternative splicing at the single-molecule level remain poorly characterized — because established methods (MeRIP-seq, m⁶A-CLIP) measure these features in separate experiments at population-averaged resolution. Kim and colleagues applied Nanopore direct RNA sequencing (dRNA-seq) to the MOLM13 human leukemia cell line with and without METTL3 knockdown, generating a multi-feature epitranscriptome dataset that simultaneously captured m⁶A status, poly(A) tail length, isoform structure, and RNA abundance from the same native RNA molecules — providing the first integrated single-molecule view of how these regulatory features interact across the human transcriptome.

2. Methods

PolyA⁺ RNA was extracted from MOLM13 cells (control and METTL3 knockdown, triplicate each) and sequenced on ONT PromethION using direct RNA sequencing chemistry, producing over 34 million reads across 6 samples. m⁶A modifications were called using m6Anet (multiple-instance learning) at single-nucleotide resolution, with per-read probability scores enabling both site-level and molecule-level modification analysis. Poly(A) tail lengths were estimated from the raw signal using Nanopolish, providing per-read tail length measurements for every captured transcript. Isoform quantification was performed using the dRNA-seq data to assess alternative splicing and isoform-specific m⁶A patterns. Correlation analyses between m⁶A status, poly(A) tail length, RNA abundance, and alternative splicing were performed within individual reads across the transcriptome.

3. Results

Nanopore direct RNA sequencing of MOLM13 leukemia cells — simultaneous m6A, poly(A) tail, and isoform analysis from Kim et al. 2025 Figure 3. Multi-feature epitranscriptome analysis by direct RNA sequencing. (A) Schematic of the dRNA-seq workflow showing simultaneous capture of m⁶A modification, poly(A) tail length, isoform structure, and RNA abundance from native RNA molecules. (B) m⁶A modification profiles across MOLM13 transcripts with METTL3 knockdown effect. (C) Correlation analysis of m⁶A status, poly(A) tail length, and mRNA abundance at single-molecule resolution. From Kim et al. (2025, Cell Genomics, CC BY 4.0).

Key Findings

  • ~25,000 m⁶A sites identified per replicate at single-nucleotide resolution — the study produced the most comprehensive single-molecule m⁶A catalog from dRNA-seq in a human leukemia cell line, with per-read modification probability scores enabling isoform-level and allele-specific m⁶A analysis that is structurally impossible from MeRIP-seq peak data
  • m⁶A modification showed a negative correlation with mRNA abundance across the transcriptome — transcripts with higher m⁶A levels tended to have lower abundance, consistent with the established role of m⁶A in promoting mRNA decay, but with substantial transcript-to-transcript variation that was captured only because the dRNA-seq data reported modification status and abundance on the same native RNA molecules
  • Poly(A) tail length was negatively associated with both mRNA abundance and stability — challenging the traditional model that longer poly(A) tails uniformly increase mRNA stability, and revealing a more complex relationship that depends on m⁶A modification context. This finding was enabled specifically by the dRNA-seq approach, which measures tail length and modification status on the same molecules
  • METTL3 knockdown (~5–10% global m⁶A reduction) revealed pathway-specific epitranscriptomic effects — m⁶A loss disproportionately affected transcripts in RNA processing, translation, and mitochondrial function pathways, with coordinated changes in poly(A) tail length and alternative splicing that varied by gene and isoform
  • Isoform-level m⁶A patterns differed substantially within multi-isoform genes — transcript isoforms from the same gene carried systematically different m⁶A profiles, with some isoforms showing m⁶A enrichment in 3′ UTRs while alternative isoforms from the same gene lacked these modifications, revealing a layer of epitranscriptomic regulation that is completely invisible when modifications are mapped to collapsed gene models from fragmented short-read data

4. Conclusions

This study provides direct experimental validation that Nanopore direct RNA sequencing delivers biologically meaningful, single-molecule m⁶A modification data simultaneously with poly(A) tail length, isoform structure, and RNA abundance from the same native RNA molecules — without antibody enrichment, chemical conversion, or parallel experiments. The findings demonstrate that relationships between m⁶A modification, poly(A) tail length, and mRNA stability are more complex and context-dependent than previously understood from population-averaged methods, and that isoform-resolved m⁶A analysis reveals patterns of epitranscriptomic regulation that are structurally inaccessible to short-read or antibody-based approaches. These results directly validate our integrated dRNA-seq service design: for projects requiring comprehensive single-molecule epitranscriptome characterization — including m⁶A profiling, poly(A) tail analysis, isoform discovery, and their integration — Nanopore direct RNA sequencing currently represents the only technology that captures all four information layers from the same native RNA molecules in a single experiment.

When to Choose Long-Read RNA Methylation Sequencing — and When Alternative Methods May Be More Suitable

Choose long-read RNA methylation sequencing when:

Consider alternative methods when:

CD Genomics provides free project consultation to help determine the optimal RNA methylation analysis strategy for your specific research questions. Contact our scientists to discuss your project requirements.

Interpretation Boundaries for Long-Read RNA Methylation Detection

  • RNA modification calls from dRNA-seq are probabilistic, not deterministic. Modification status at each nucleotide is reported as a probability score derived from neural-network-based basecalling models or comparative signal analysis. Per-read modification calls at individual sites become statistically robust at ≥20× coverage per site. At lower coverage, per-site aggregate modification frequency (fraction of modified reads) is more reliable than individual read-level calls. We report both per-read and per-site metrics to support appropriate interpretation
  • Modification detection sensitivity varies by modification type, sequence context, and coverage. m⁶A detection in DRACH motif contexts achieves the highest sensitivity and specificity, while detection outside canonical motifs or of less common modifications (pseudouridine, inosine) may require higher coverage or matched unmodified (IVT) control libraries for optimal performance. Detection rates for each modification type are documented in project reports
  • dRNA-seq data are for research use only. All RNA modification calls, isoform annotations, and differential modification analyses are generated for research applications and are not validated for clinical diagnosis, treatment decisions, or clinical variant interpretation in individual patients or specimens
  • RNA integrity at the time of extraction determines data quality. Degraded RNA produces truncated reads that reduce full-length transcript coverage and modification detection sensitivity, particularly for modification sites in the 5′ regions of long transcripts. RNA quality is assessed before library preparation and documented in the project QC report
  • Differential modification analysis requires appropriate statistical power. Identifying statistically significant modification differences between conditions requires adequate biological replication (minimum 3 replicates per condition recommended) and coverage depth per site. False discovery rate control is applied to all differential modification calls. We provide power analysis guidance during project design

FAQs

RNA Methylation Analysis Deliverables Include Per-Site Modification Frequencies, Single-Molecule Profiles, and Integrated Epitranscriptome Reports

1. Per-site modification frequency table (BEDmethyl/MM/ML SAM tag format) with modification probability scores for every detected modification site in the transcriptome

2. Per-read single-molecule modification status file enabling isoform-resolved, allele-specific, and co-occurrence modification analysis

3. Differential modification analysis report (per-site and per-isoform) with statistical significance, effect sizes, and genomic annotation across experimental groups

4. Transcriptome browser tracks (bigWig, BED) for modification probability visualization in IGV or UCSC Genome Browser, including per-modification-type tracks (m⁶A, m⁵C, Ψ) and isoform-level modification profiles

5. Optional: integrated multi-feature report combining m⁶A modification, poly(A) tail length distribution, isoform quantification, and cross-feature correlation analysis with publication-ready figures — based on the same analytical framework validated in Kim et al. (2025, Cell Genomics)

Sample RNA methylation analysis report showing per-site m6A modification frequencies, poly(A) tail length distribution, and isoform-level modification profiles

For research use only. Not for use in diagnostic procedures.

Get Your Instant Quote