m⁵C (5-methylcytosine) is a common RNA modification found across mRNA, tRNA, rRNA, and many non-coding RNAs. It has been linked to RNA stability, translation, and RNA–protein interactions. Considering this, it's no surprise that mapping m⁵C has become a widespread project goal in recent years.

However, mapping m⁵C is not without its challenges. Many projects go off-track at the critical stage of method–question selection. Specifically, an enrichment map (peaks) is not indicative of exact cytosines (sites). And "sites" without a defensible methylation fraction (stoichiometry) can be hard to interpret across conditions.

This article helps you choose a method that matches the claim you want to make, then concisely explains the trade-offs. Real research examples are given for each method so your team can see how it's applied in practice.

If you're new to epitranscriptomics and want a quick orientation before choosing an assay, see What is RNA Methylation and How to Study.

All workflows discussed here are for research use only (RUO) and are not intended for diagnostic or treatment decisions.

Questions to Start

"How Much" or "Where"?

If your primary goal is to quantify whether m⁵C changes across conditions, LC–MS/MS provides a direct abundance readout at the nucleoside level. If you need to locate changes on transcripts, you'll need a sequencing-based method.

Do You Require a Screen, or Base-level Evidence?

RIP-seq is suited for fast transcriptome-wide screening and reports enriched regions (peaks). Base-resolution approaches (BS-m⁵C-seq and UBS-seq) support site-level calls and enable more interpretable and easier to justify comparisons at specific cytosines. UBS-seq is generally designed to reduce classic bisulfite-associated damage/artifacts.

Is Isoform or Single-molecule Context Relevant to Your Biology?

If interpretation depends on isoforms or co-occurrence on individual molecules, Nanopore direct RNA sequencing (DRS) can add full-length, single-molecule context through modification calling. If isoform context is not central, short-read mapping workflows are usually easier to benchmark and standardize.

For Quick Comparison

Practical Selection Guide (preferred by most labs):

• Need an overall abundance readout? Often start with LC–MS/MS.
• Need transcriptome-wide discovery at base resolution? Consider UBS-seq (or carefully controlled BS-m⁵C-seq when needed).
• Need quick exploratory maps/regions? m⁵C-RIP/MeRIP-seq is typically faster but reports peaks, not exact sites.
• Want isoform-resolved, single-molecule context? Consider Nanopore DRS + mod-calling, with strong controls and benchmarking.

The table below summarizes what each method can (and cannot) claim, along with practical trade-offs in input, artifacts, and interpretability.

Compare methods at a glance

For Quick Decisions

Contact us with:

(1) Your RNA type (total RNA / mRNA-enriched / tRNA/rRNA), (2) input amount, and (3) what you need to report (peaks vs sites vs fractions vs isoform context).

We'll recommend the best-fit method (LC–MS/MS, RIP-seq, BS-m⁵C-seq, UBS-seq, or Nanopore DRS) and a QC plan you can defend.

Talk to a scientist | Request a quote

LC–MS/MS (mass spectrometry, nucleoside quantification)

LC–MS/MS quantifies m⁵C at the nucleoside level following enzymatic digestion, providing a sensitive readout of overall m⁵C abundance in a defined RNA pool. However, it does not localize sites on transcripts.

LC-MS/MS is best suited for:

• Confirming whether m⁵C abundance shifts across conditions (e.g., KO/KD, treatment, stress)
• Perturbation studies of m⁵C writers/readers where orthogonal quantitative checks are desired
• QC and cross-validation alongside mapping approaches (RIP-seq / BS-m⁵C-seq / UBS-seq)
• Measurements focused on a specific RNA fraction (e.g., <200 nt small RNAs enriched for tRNAs)

Typical workflow (high-level)

• RNA extraction and QC
• Optional fractionation/enrichment (e.g., small RNA, poly(A)+), aligned with the biological question
• Enzymatic digestion to single nucleosides (plus dephosphorylation as required)
• Internal standards + LC–MS/MS acquisition (often MRM on triple quadrupole)
• Quantification and normalization (with batch/QC review)

Key considerations

• RNA pool must be defined up front (total RNA vs poly(A)+ vs small RNA vs purified tRNA can yield very different readouts)
• A control is required for carryover/contamination between fractions (and for residual DNA if relevant to your extraction)
• Digestion completeness and instrument QC strongly affect comparability; technical replicates and process controls can help
• LC–MS/MS is quantitative but not positional—pair with sequencing when transcript/site localization is needed

Typical deliverables

• Per-sample table reporting m⁵C abundance for the specified RNA pool (with clear normalization)
• QC summary (replicate agreement, internal standards performance, process controls)
• Optional stratified reporting by RNA fraction (if multiple fractions are tested)

Study example

This study used LC–MS/MS to quantify RNA m⁵C levels in small RNA/tRNA-enriched pools, demonstrating how LC–MS/MS can serve as a quantitative orthogonal readout for genetic perturbations of m⁵C writers such as NSUN2.

Source: Gonskikh Y, Tirrito C, Bommisetti P, et al. Spatial regulation of NSUN2-mediated tRNA m⁵C installation in cognitive function. Nucleic Acids Research (2025). doi:10.1093/nar/gkae1169.

Background: NSUN2 is a major m⁵C methyltransferase for cytoplasmic tRNAs, and disease-linked variants can impair m⁵C installation. A quantitative assay is helpful for comparing global changes in tRNA-associated m⁵C across knockout and rescue conditions.

Research methods: The study performed LC–MS/MS analyses on small RNAs (<200 nt), which are largely tRNA-enriched, to quantify m⁵C levels in NSUN2 knockout cells and after re-expression of wild-type or mutant NSUN2.

Research results: LC–MS/MS showed a significant reduction of m⁵C levels in small RNAs upon NSUN2 knockout, and demonstrated that wild-type NSUN2 (but not the G679R disease mutant) restored m⁵C levels in the small RNA pool, supporting the conclusion that the mutant is deficient in installing m⁵C.

m⁵C-RIP-seq (antibody enrichment)

m⁵C-RIP-seq is an antibody-enrichment sequencing approach that identifies transcriptome-wide RNA regions enriched for m⁵C signal. It is well-suited for discovery screening but does not pinpoint exact cytosine sites.

RIP-seq is best suited for:

• Transcriptome-wide screening across specific conditions (e.g., treatment vs control)
• Evaluating/prioritizing candidate transcripts/regions for downstream site-level validation
• Exploring global enrichment patterns when base-resolution coordinates are not required

Typical workflow (high-level)

1. RNA extraction and QC
2. RNA fragmentation
3. m⁵C antibody immunoprecipitation (IP) and Input preparation
4. Library preparation (IP + Input)
5. Sequencing and enrichment/peak analysis

Key considerations

• Results are reported as enriched regions/peaks, not definitive single-base sites.
• Data quality is sensitive to antibody performance and replicate consistency.
• Input controls are essential for interpreting enrichment versus background.

For a practical primer on RIP-seq experimental design and data analysis, read our Comprehensive Guide to RIP-Seq Technology.

Typical deliverables

• Peak/enrichment region list with transcript annotation
• Genome browser tracks (IP and Input)
• Differential enrichment analysis (when study design includes replicates)

Study example

This study reported that aberrant RNA m⁵C hypermethylation promotes intrinsic gefitinib resistance in EGFR-mutant NSCLC through an NSUN2–YBX1–QSOX1 regulatory axis.

Source: Wang Y, Wei J, Feng L, et al. Aberrant m5C hypermethylation mediates intrinsic resistance to gefitinib through NSUN2/YBX1/QSOX1 axis in EGFR-mutant non-small-cell lung cancer. Molecular Cancer (2023). doi:10.1186/s12943-023-01780-4.

Background: Intrinsic resistance to EGFR-TKIs remains a major cause of treatment failure in EGFR-mutant NSCLC, and the role of RNA m⁵C modification in tumor drug resistance was unclear.

Research methods: To assess the correlation between RNA m⁵C methylation, the m⁵C writer NSUN2, and EGFR-TKI resistance in NSCLC cell lines and patient samples, gain- and loss-of-function assays were performed in vitro and in vivo. Additionally, integrated RNA-seq, RNA bisulfite sequencing (RNA-BisSeq), and MeRIP-qPCR were used to identify downstream targets and mechanisms.

Research results: RNA m⁵C hypermethylation and NSUN2 were correlated with intrinsic EGFR-TKI resistance. NSUN2 overexpression induced gefitinib resistance and tumor recurrence, while NSUN2 inhibition overcame intrinsic resistance in vitro and in vivo. Integrated RNA-seq and RNA-BisSeq analyses nominated QSOX1 as a target, and NSUN2-mediated methylation in the QSOX1 coding region enhanced QSOX1 translation via the m⁵C reader YBX1.

CD Genomics provides services for research use only (RUO) and does not offer clinical testing or medical advice.

BS-m⁵C-seq (RNA bisulfite sequencing)

BS-m⁵C-seq uses bisulfite conversion to yield unmodified cytosines that read as T (while m⁵C remains C), enabling single-base m⁵C site calling and methylation-fraction estimation when coverage and QC are sufficient.

BS-m⁵C-seq is best suited for:

• Single-base m⁵C site mapping with coordinate-level reporting
• Comparing methylation fraction (stoichiometry) across groups (coverage/QC dependent)
• Mechanistic studies requiring precise sites for follow-up (e.g., mutagenesis, reader binding)

Typical workflow (high-level)

1. RNA extraction and QC
2. Fragmentation and structure handling (as needed)
3. Bisulfite conversion
4. Reverse transcription and library construction
5. Sequencing, alignment, site calling, and fraction estimation

Key considerations

• Bisulfite treatment can induce RNA degradation, which impacts coverage—especially for low-abundance transcripts.
• C-to-U/T conversion reduces read complexity, so mapping strategy and filtering thresholds strongly affect final site lists.
• Interpretable stoichiometry requires adequate coverage and transparent QC criteria.

Typical deliverables

• Base-resolution site list with methylation fraction and coverage
• QC summary (mapping metrics, duplication, conversion/background indicators)
• Summary plots (distribution across transcript features; optional context summaries)

Study example

This study showed that loss of NSUN5 reduces m⁵C and disrupts maternal mRNA stability during the maternal-to-zygotic transition. This is associated with altered ovarian function in the experimental system and arrested embryonic development.

Source: Ding C, Lu J, Li J, et al. RNA-methyltransferase Nsun5 controls the maternal-to-zygotic transition by regulating maternal mRNA stability. Clinical and Translational Medicine (2022). doi:10.1002/ctm2.1137.

Background: m⁵C-mediated regulation of mRNA decay and stability is critical during maternal-to-zygotic transitioning (MZT). The contribution of NSUN5-related m⁵C changes to oogenesis, ovarian function, and embryonic development warranted clarification, which this study offered.

Research methods: Mouse ovaries and oocytes with Nsun5 knockout (KO) and the KGN cell line were evaluated using m⁵C identification approaches (including BS-m⁵C-seq), alongside analyses of alternative splicing and protein expression.

Research results: The study reported that NSUN5 deletion inhibited ovarian function and led to arrested embryonic development, and BS-m⁵C-seq data supported a model in which reduced m⁵C levels following NSUN5 knockout were associated with time-dependent acceleration of ovarian aging; NSUN5 loss was also linked to decreased MAD2L2 and GDF9 protein levels.

UBS-seq (ultrafast bisulfite sequencing)

UBS-seq is an ultrafast bisulfite-based workflow designed to generate single-base m⁵C sites and methylation fractions with a faster conversion step and improved practicality for challenging or low-input samples compared with standard bisulfite protocols.

UBS-seq is best suited for:

• Base-resolution mapping when input is limited or sample integrity is a concern
• Projects where reducing workflow time and background helps improve practical robustness
• Studies that still require sites + methylation fraction, but want a more operationally efficient conversion step

Typical workflow (high-level)

1. Sample preparation (RNA; matched DNA methylation assays can be designed separately if needed)
2. Ultrafast conversion under optimized conditions
3. Cleanup/desulfonation
4. Library construction
5. Sequencing, alignment, site calling, and fraction estimation with strict QC/filters

Key considerations

• As with all conversion-based approaches, mapping and filtering remains critical due to reduced sequence complexity after conversion.
• Tight control of reaction conditions is important for reproducibility.
• Outputs are intended for research use and should be interpreted within study design and QC constraints.

Typical deliverables

• List of base-resolution sites with methylation fraction and coverage
• QC package emphasizing background/non-conversion and processing consistency
• Optional integrated interpretation with matched RNA-seq / phenotype data (project-dependent)

Study example

This study reported on Ultrafast Bisulfite Sequencing (UBS-seq), an accelerated bisulfite conversion strategy that enables quantitative RNA m⁵C mapping from low-input mRNA and supports estimating m⁵C stoichiometry in highly structured RNAs.

Source: Dai Q, Ye C, Irkliyenko I, et al. Ultrafast bisulfite sequencing detection of 5-methylcytosine in DNA and RNA. Nature Biotechnology (2024). doi:10.1038/s41587-023-02034-w.

Background: RNA bisulfite sequencing can be confounded by RNA degradation and structure-associated incomplete C-to-U conversion, which increases background and false positives—issues that become more apparent when profiling low-abundance RNA such as mRNA.

Research methods: Researchers successfully shortened the conversion window by using concentrated bisulfite chemistry and high-temperature conversion. Then, they applied UBS-seq to poly(A)+ mRNA (including HeLa and HEK293T) and benchmark RNA m⁵C detection with perturbations of known writers (notably NSUN2 and a smaller NSUN6 component) to evaluate signal specificity and quantitative behavior.

Research results: This study demonstrated that UBS-seq supports quantitative mapping of RNA m⁵C from low-input mRNA, and enables stoichiometry measurement even in highly structured RNA sequences. The study identifies NSUN2 as the major writer responsible for ～90% of m⁵C sites in HeLa mRNA and reports enrichment of m⁵C sites toward 5′ regions of mammalian mRNAs.

Nanopore DRS (Oxford Nanopore direct RNA sequencing)

Nanopore direct RNA sequencing (DRS) reads native RNA molecules and preserves raw electrical signal features that can be leveraged for modification inference, adding isoform- and single-molecule–aware context; interpretation typically depends on controls and benchmarking.

Nanopore DRS is best suited for:

• Isoform-resolved questions where full-length context changes interpretation (e.g., isoform-specific regulation)
• Designs that can provide a matched "lower-modification" control (e.g., writer KO/KD, IVT RNA) for comparative calling
• Cases where sequence-isoforms-candidate modification signals are desired from one assay, followed by focused validation

Typical workflow (high-level)

• RNA extraction and QC (often poly(A)+ input for standard DRS workflows)
• DRS library preparation and sequencing
• Basecalling and alignment to reference
• Signal-to-reference alignment (event-level processing) for modification-aware analyses
• Modification inference (comparative or model-based) + filtering/QC
• Orthogonal validation for prioritized candidates (recommended for key claims)

Key considerations

• Controls are central: comparative approaches explicitly rely on a "less modified" reference; model-based callers depend on training/benchmarking
• Coverage and expression level constrain sensitivity; low-abundance transcripts may remain ambiguous
• Toolchain versions (chemistry/basecaller/model) can change performance; be sure to report versions and keep thresholds consistent
• Treat calls as candidates unless supported by strong controls and validation, especially when making site-level statements

Typical deliverables

• FASTQ + alignments (BAM), basic yield/read-length/alignment QC
• Transcript/isoform summaries (where applicable)
• Candidate modification calls with confidence metrics (and differential signals for comparative designs)
• A prioritized candidate list for follow-up validation

Study example

This study introduced Nanocompore, a comparative analysis framework for detecting RNA modification signals from Nanopore direct RNA sequencing (DRS). This was achieved by contrasting a sample of interest against a matched "non-modified or modification-depleted" control, enabling replicate-aware statistical calling in signal space.

Source: Leger A, Amaral PP, Pandolfini L, et al. RNA modifications detection by comparative Nanopore direct RNA sequencing. Nature Communications (2021). doi:10.1038/s41467-021-27393-3.

Background: Nanopore DRS reads native RNA, procuring a raw electrical signal that can be perturbed by nucleotide modifications. However, robust detection is challenging because signal changes depend on sequence context, coverage, and technical variability. A practical approach is to compare modified versus non-modified (or depleted) conditions to attribute consistent signal shifts to modifications.

Research methods: Nanocompore applies a model-free comparative strategy that aggregates per-read signal features (including intensity and dwell time) at transcript-position level. This is followed by tests for distributional differences between conditions, using univariate tests and a bivariate mixture-model approach; the workflow is designed to use replicates and does not require a training set.

Research results: This study demonstrated that the comparative framework can detect multiple RNA modification types in vitro with positional accuracy. The framework was applied to profile modifications in biological RNAs and supported interpretation with orthogonal confirmation. This highlights the importance of matched controls and validation when translating DRS signals into modification candidates.

For applied use cases of DRS in RNA modification research, see ONT Direct RNA Sequencing: Translational Applications in RNA Modification Studies.

Common Project Paths

Path A: Verify whether m⁵C changes at all, then decide how deep to investigate.

Start with LC–MS/MS when you need an absolute, chemistry-based readout of overall m⁵C abundance across a defined RNA fraction. If you see a clear shift, move to site/region mapping (UBS-seq or BS-m⁵C-seq) or a faster enrichment screen (RIP-seq) depending on resolution needs.

Path B: Fast transcriptome-wide screening, then pinpoint.

Use RIP-seq to obtain enriched regions/peaks and shortlist candidates, then validate key loci with UBS-seq (preferred when you want reduced damage/background) or BS-m⁵C-seq (when you need classic bisulfite-compatible single-base calls and can manage artifacts with controls).

Path C: Base-resolution discovery first (publication-driven).

Go directly to UBS-seq / BS-m⁵C-seq when you need transcriptome-wide, single-base candidates. Build in strict conversion/coverage QC and plan orthogonal checks (e.g., targeted validation; LC–MS/MS can serve as a global orthogonal measurement when interpretation is controversial).

Path D: Isoform context or single-molecule co-occurrence matters.

Choose Nanopore DRS when isoform-resolved context, long-range co-occurrence, or multi-modification potential is central. Treat mod-calling as a comparative/controlled inference step and validate high-confidence candidates with an orthogonal assay when needed.

Notes for Study Design

• Replicates: If you plan to compare conditions, biological replicates are the critical difference between "suggestive" and "defensible."
• Varying controls: RIP-seq requires Input (and optional IgG) to interpret enrichment; conversion workflows need pre-declared conversion/background QC and coverage thresholds; LC–MS/MS should use stable standards/process controls and a consistent RNA fraction; Nanopore DRS benefits from matched comparative controls and fixed toolchain versions across conditions.
• Define what you will report: peaks (regions) vs sites (coordinates) vs fractions (stoichiometry) vs isoform/single-molecule context.
• Plan for failure modes: understand what to do if coverage is low, replicate concordance is weak, or mapping drops after conversion.

Conclusion

RNA m⁵C detection is no longer a one-method decision, as different technologies answer to different questions. LC–MS/MS provides a chemistry-based abundance readout for a defined RNA fraction, RIP-seq offers rapid enrichment-based screening, BS-m⁵C-seq and UBS-seq support base-resolution candidate discovery with different artifact profiles, and Nanopore DRS adds isoform and single-molecule context while requiring careful controls and benchmarking. In practice, many projects combine a fast screen with a higher-resolution confirmation step to balance interpretability, cost, and turnaround.

CD Genomics services for m⁵C profiling

Our team supports RNA m⁵C studies using LC–MS/MS, m⁵C-RIP/MeRIP-seq, BS-m⁵C-seq, UBS-seq, and Nanopore DRS + mod-calling, covering abundance measurement, enrichment screening, base-resolution mapping, and isoform/single-molecule context. We help researchers like you select the best-fit workflow based on your question, sample constraints, and reporting needs. Research use only (RUO).

If you'd like to discuss your study design and sample type, feel free to reach out for a project consultation and quotation. Deliverables can include raw data, QC summaries, and peak/site/fraction/candidate call lists aligned to your reporting goal.

FAQ

Can m⁵C-RIP-seq pinpoint the exact modified cytosine?

No. m⁵C-RIP-seq reports enriched regions, not a definitive base-level coordinate. It's best used for screening and prioritization, then followed by base-resolution methods if exact sites are needed.

What does "methylation fraction/stoichiometry" mean in m⁵C sequencing?

This figure represents the fraction of molecules methylated at a specific cytosine in your sample. In practice, it's only interpretable when you have adequate coverage, low background non-conversion, and replicate concordance.

Why do different studies report different mRNA m⁵C site lists?

Because conversion-based calling is sensitive to RNA integrity, mapping complexity after conversion, coverage thresholds, and filtering rules. Small differences in any of these can shift low-fraction, low-coverage sites in or out of the final list.

When should I choose UBS-seq over standard RNA bisulfite sequencing?

Choose UBS-seq when you still need base-resolution sites and fractions, but your constraints make standard bisulfite risky—especially low input, fragile samples, or cases where background noise and processing loss are likely to dominate.

Which method is best if I only need a shortlist of candidate genes?

m⁵C-RIP-seq is usually the most direct route for shortlist generation, as long as you treat peaks as regions of interest rather than exact sites.

Can LC–MS/MS tell which transcripts or sites carry m⁵C?

No. LC–MS/MS provides a chemistry-based measurement of m⁵C abundance for a defined RNA fraction, but it does not localize modifications to specific transcripts or cytosines. Use sequencing-based approaches (UBS-seq / BS-m⁵C-seq) or controlled long-read inference (Nanopore DRS) when location matters.

Does Nanopore DRS provide definitive single-base m⁵C calls?

DRS preserves native RNA and can support isoform-resolved, single-molecule analyses, but modification calling depends on comparative design, signal models, and coverage. For publication-grade claims, prioritize replicate-supported candidates and validate key findings with an orthogonal assay when appropriate.

References

1. Wang Y, Wei J, Feng L, et al. Aberrant m5C hypermethylation mediates intrinsic resistance to gefitinib through NSUN2/YBX1/QSOX1 axis in EGFR-mutant non-small-cell lung cancer. Molecular Cancer. 2023.
2. Ding C, Lu J, Li J, et al. RNA-methyltransferase Nsun5 controls the maternal-to-zygotic transition by regulating maternal mRNA stability. Clinical and Translational Medicine. 2022.
3. Dai Q, Ye C, Irkliyenko I, et al. Ultrafast bisulfite sequencing detection of 5-methylcytosine in DNA and RNA. Nature Biotechnology. 2024.
4. Khoddami V, Cairns BR. Identification of direct targets and modified bases of RNA cytosine methyltransferases. Nature Biotechnology. 2013.
5. Yang X, Yang Y, Sun BF, et al. 5-methylcytosine promotes mRNA export — NSUN2 as the methyltransferase and ALYREF as an m5C reader. Cell Research. 2017.
6. Leger A, Amaral PP, Pandolfini L, et al. RNA modifications detection by comparative Nanopore direct RNA sequencing. Nature Communications. 2021.
7. Gonskikh Y, et al. Spatial regulation of NSUN2-mediated tRNA m5C installation in cognitive function. Nucleic Acids Research. 2025.
8. Hussain S, Sajini AA, Blanco S, et al. NSun2-mediated cytosine-5 methylation of vault noncoding RNA determines its processing into regulatory small RNAs. Cell Reports. 2013.
9. Squires JE, Patel HR, Nousch M, et al. Widespread occurrence of 5-methylcytosine in human coding and non-coding RNA. Nucleic Acids Research. 2012.
10. Amort T, Rieder D, Wille A, et al. Distinct 5-methylcytosine profiles in poly(A) RNA from mouse embryonic stem cells and brain. Genome Biology. 2017.
11. Legrand C, Tuorto F, Hartmann M, et al. Statistically robust methylation calling for whole-transcriptome bisulfite sequencing reveals distinct methylation patterns for mouse RNAs. Genome Research. 2017.
12. Lu L, Zhang X, Zhou Y, et al. Base-resolution m5C profiling across the mammalian transcriptome by bisulfite-free enzyme-assisted chemical labeling approach. Molecular Cell. 2024.