m6A research is moving from cataloging sites to measuring modification levels. The early epitranscriptomics workflow — antibody enrichment, peak calling, gene-level annotation — answered a foundational question: where is m6A in the transcriptome? That question produced landmark datasets from methods such as MeRIP-seq and miCLIP, which mapped tens of thousands of m6A sites across human and mouse transcriptomes. Those maps remain valuable references, but the questions researchers ask today have shifted.

The current frontier is not where m6A is, but how much m6A is present at each site, how that amount differs between conditions, and how modification level relates to function. Answering those questions requires quantification — an absolute or near-absolute measurement of the fraction of transcripts carrying m6A at a given position. Detection tells you the site exists. Quantification tells you whether 10% or 80% of transcripts are modified there, and whether that proportion changes under experimental perturbation.

This article describes the difference between detection and quantification in m6A analysis, identifies the biological scenarios where detection alone is insufficient, and provides a framework for deciding when to adopt a quantitative method such as GLORI-seq.

Figure 1: Detection reports whether an m6A site exists; quantification measures the fraction of transcripts modified at that site.

What Detection Delivers — And Where It Stops

m6A detection methods share a common output: they report the genomic location of m6A modifications. Antibody-based methods such as MeRIP-seq enrich methylated RNA fragments and identify peaks of enrichment relative to input RNA. miCLIP achieves higher resolution by crosslinking the antibody to the modification site, producing characteristic truncation or mutation signatures at or near the modified nucleotide. Enzymatic methods such as DART-seq produce C-to-U edits at positions adjacent to m6A residues, marking modification sites through an indirect editing signature.

These methods answered the first-order question of epitranscriptomics — cataloging the m6A landscape — with considerable success. The Dominissini and Meyer laboratories each published transcriptome-wide m6A maps in 2012 that defined the distribution of m6A across gene features and identified the RRACH consensus motif. Those datasets remain among the most cited resources in the field.

But detection methods share an inherent limitation: the signal they produce does not correspond directly to the fraction of transcripts modified. In MeRIP-seq, peak height is influenced by RNA abundance, antibody efficiency, and local sequence context in addition to m6A occupancy. In DART-seq, the C-to-U editing rate reflects APOBEC1 activity and fusion protein expression as much as it reflects m6A stoichiometry. A tall peak or a frequent edit does not mean a high modification level — it means the site is detectable.

This creates a gap between the data these methods produce and the questions researchers increasingly want to ask. Knowing that a site is modified is useful for discovery. Knowing what fraction of transcripts carries the modification is necessary for comparison.

What Quantification Adds

Quantitative m6A measurement provides three pieces of information that detection alone cannot.

Stoichiometry. The modification fraction — what percentage of transcripts carry m6A at a given position — is the fundamental unit of quantitative m6A analysis. Stoichiometry ranges from 0 (no transcripts modified) to 1 (all transcripts modified), and real biological sites span this entire range. A site modified at 10% of transcripts and a site modified at 80% cannot be distinguished by detection methods; both produce a “present” call. But the two sites almost certainly have different biological roles, different sensitivities to writer concentration, and different functional consequences.

Comparability across conditions. Quantitative methods produce modification fractions that are directly comparable between treatment and control, wild-type and knockout, or disease and healthy samples. A shift from 0.3 to 0.5 at a given site is interpretable as a change in modification level, independent of whether the transcript itself changed in abundance. Detection methods report whether a site was “found” in each condition — which conflates changes in modification with changes in expression.

Comparability across sites. Stoichiometric measurements allow comparison of modification levels between different sites in the same transcript, or between sites in different transcripts. This enables questions such as “does the first m6A site in the 3′ UTR carry a higher modification level than the second?” or “are sites near the stop codon modified at higher stoichiometry than sites further downstream?” — questions that are unanswerable with binary detection data.

These three properties — stoichiometry, cross-condition comparability, and cross-site comparability — are what make quantitative m6A data suitable for mechanistic inference. They are also the properties that the best-established quantitative method, GLORI-seq, was designed to provide.

The Biological Questions That Require Quantification

Not every m6A project requires quantification. A project designed to catalog m6A sites in a previously uncharacterized organism or tissue may be well served by a detection method. But several classes of biological questions are difficult or impossible to address without quantitative modification data.

Condition comparison. “Is m6A higher in treatment than in control at this site?” This is the most common differential analysis question, and it requires a quantitative readout. Detection methods can report whether a site was found in each condition, but a site that is detected in treatment and not in control could reflect a genuine gain of m6A, higher transcript expression in treatment, or a technical difference in IP efficiency between samples. Quantitative methods disambiguate these possibilities by reporting the modification fraction independently of expression.

Time-course and dynamic studies. m6A is a dynamic modification that changes in response to stimuli such as heat shock, hypoxia, and nutrient deprivation. Tracking these changes over time requires measuring modification level at each time point, not simply whether a site is present or absent. A site that shifts from 20% to 40% modification after stimulation represents a biologically meaningful doubling of m6A; detection methods would report “present” at both time points and miss the change entirely.

Writer and eraser perturbation. Experiments that knock down or knock out m6A writers (METTL3, METTL14, WTAP) or erasers (FTO, ALKBH5) aim to identify sites whose modification depends on a specific enzyme. The biologically informative output is not whether sites disappear — complete loss of m6A at a site is rare even in writer knockouts — but whether modification level decreases. A site that drops from 60% to 30% modification upon METTL3 depletion is a strong candidate for a METTL3 target. Detection methods that report a binary present/absent call cannot measure these partial reductions, which represent the most common outcome of writer perturbation.

Stoichiometry-function relationships. A growing body of evidence suggests that m6A function depends on modification level, not just modification presence. High-stoichiometry sites may have different effects on RNA stability, translation, or structure than low-stoichiometry sites. Testing these relationships requires measuring stoichiometry, not just detecting the modification. This is particularly relevant for studies examining how m6A affects RNA secondary structure, protein binding, or transcript localization — processes that may respond to the fraction of modified transcripts in a graded rather than binary manner.

For each of these scenarios, the limitation is the same: detection data can tell you that a difference exists, but it cannot tell you how large the difference is, whether it is independent of expression, or whether it follows a dose-response relationship with the perturbation.

Methods That Quantify m6A at Single-Base Resolution

Several methods have been developed that move beyond detection to provide quantitative m6A measurement. They differ in chemistry, resolution, and the type of quantification they provide.

GLORI-seq. GLORI-seq uses glyoxal and nitrite to chemically deaminate unmethylated adenosines to inosines, which are read as guanosines during sequencing. m6A residues are protected by the methyl group and remain read as adenosine. The fraction of reads retaining A at each position directly reports the m6A stoichiometry at that site. Because the readout is at the modification site itself — not at an adjacent position — the measurement is direct and the resolution is genuinely single-base. The original publication identified over 210,000 m6A sites in mammalian transcriptomes and reported absolute modification fractions for each. For a step-by-step walkthrough of the chemistry and protocol, see our guide to GLORI-seq principles and experimental steps.

m6A-SAC-seq. This method uses an enzymatic approach — the MazF endoribonuclease, which cleaves at unmethylated ACA motifs but not at m6A-modified ACA — combined with adaptor ligation to quantify modification levels at ACA-containing sites. It provides site-level quantification but is restricted to sites within the ACA motif context.

eTAM-seq. Enzyme-assisted adenosine deamination followed by sequencing uses an evolved TadA deaminase to convert unmethylated A to I. Like GLORI-seq, it produces a quantitative readout at single-base resolution, though with different enzyme-specific biases and substrate preferences.

m6A-REF-seq. This method combines m6A-sensitive RNA endoribonuclease cleavage with sequencing to quantify modification at specific motif sites. It resolves m6A stoichiometry at targeted positions but has a narrower transcriptome-wide coverage than GLORI-seq.

The key distinction across these methods is between absolute and relative quantification. Absolute quantification reports the fraction of transcripts modified at each site. Relative quantification reports whether modification differs between conditions but does not provide the baseline stoichiometry. GLORI-seq provides the closest approximation to absolute quantification currently available for transcriptome-wide m6A analysis.

For a broader method comparison across detection and quantification technologies, see our guide to choosing an m6A mapping method.

Figure 2: m6A methods differ in both resolution and quantification capability; the combination of single-base resolution and stoichiometric quantification is unique to chemistry-based methods such as GLORI-seq.

When Detection Is Still the Right Choice

Quantitative m6A methods are not always the best choice. Several scenarios favor detection methods, and recognizing them avoids unnecessary cost and complexity.

Initial surveys of uncharacterized systems. If m6A has not been mapped in your organism, tissue, or cell type, a detection method provides a cost-effective first-pass catalog. Once the landscape is mapped, quantitative follow-up on the most interesting sites or conditions may be warranted.

Severely input-limited samples. Quantitative methods such as GLORI-seq typically require approximately 1 μg of poly(A)-selected RNA per sample, though lower-input protocols are under development. When RNA is limited to nanogram quantities — sorted cell populations, early embryos, microdissected tissue — detection methods with lower input requirements may be the only feasible option. DART-seq, for example, works from 10 ng of total RNA.

Long-read isoform analysis. If the primary question involves m6A patterns on individual transcript isoforms, long-read sequencing platforms may be necessary. Long-read DART-seq has been demonstrated on the PacBio platform and can reveal how m6A sites co-occur on single molecules. Long-read GLORI-seq has not yet been established as a standard workflow.

Rapid screening. When the goal is to screen multiple conditions, time points, or genotypes for large-effect differences, a detection method provides faster and less expensive results. Quantitative follow-up can then be directed to the conditions and sites where it adds the most value.

The principle is straightforward: match the method to the question. If the question is “where is m6A?”, detection is sufficient. If the question is “how much m6A, and does it change?”, quantification becomes necessary.

Practical Considerations for Adopting Quantitative m6A Methods

Researchers transitioning from detection-based to quantitative m6A analysis should anticipate several practical differences.

Experimental design. Quantitative methods reward careful experimental design. Because the output is a modification fraction with an associated confidence interval, the design must support statistical comparison. Replicates, batch randomization, and spike-in controls matter more for quantitative methods than for detection methods, where the primary output is a list of sites. For detailed guidance, see Designing a GLORI-seq Project for Differential m6A Quantification.

Sequencing depth. Quantitative methods generally require deeper sequencing than detection methods. GLORI-seq identifies approximately 80,000 sites at 50 Gb per sample and over 210,000 sites at higher depth. The additional depth is needed because the modification fraction at each site must be estimated with enough precision to support statistical comparison — not simply called as present or absent.

Analysis complexity. Quantitative m6A data requires analysis approaches that account for both the modification fraction and the read depth supporting each estimate. Models developed for differential DNA methylation analysis — beta-binomial models, logistic regression — are often adapted for this purpose. The analysis is more computationally intensive than peak calling, but the outputs — site-level modification fractions with confidence intervals — directly support the biological questions that motivated the experiment.

Cost. Quantitative methods are generally more expensive per sample than detection methods, reflecting the additional sequencing depth and the more complex library preparation. However, the per-site information content is substantially higher. A quantitative dataset that directly answers a mechanistic question may be more cost-effective than a detection dataset that requires follow-up experiments to interpret.

For a step-by-step approach to moving from detection-based to quantitative m6A analysis, including assay selection and study design support, RNA modification services at CD Genomics provide integrated workflows spanning method consultation through data delivery.

Figure 3: A decision framework for choosing between detection and quantification methods based on the research question, sample constraints, and required outputs.

Summary

m6A research has reached a point where detection is no longer the bottleneck. Thousands of sites have been cataloged across species, tissues, and conditions. The bottleneck now is interpretation — and interpretation depends on quantification.

Detection methods answer the question “is this site modified?” Quantitative methods answer “how modified is this site, and does that change between conditions?” The second question supports mechanistic inference in a way the first does not. Stoichiometry enables comparison between conditions and between sites. Single-base resolution enables confident assignment of modification to individual adenosines rather than broad regions. Together, they provide the measurement precision that the current generation of m6A biological questions demands.

The choice between detection and quantification is not about which method is newer or more sophisticated. It is about which method produces data that can answer your question. If your question is about comparison, dynamics, or mechanism, SELECT-m6A sequencing for targeted validation and GLORI-seq for transcriptome-wide quantification are the appropriate tools. If your question is about discovery and cataloging, detection methods remain practical and appropriate. The key is to recognize which type of question you are asking before committing to the experimental design.

FAQ

1. What is the difference between m6A detection and m6A quantification?

Detection reports whether an m6A site is present — a binary yes/no call. Quantification measures the fraction of transcripts carrying m6A at each site — a continuous value between 0 and 1. Detection methods include MeRIP-seq, miCLIP, and DART-seq. Quantitative methods include GLORI-seq, m6A-SAC-seq, and eTAM-seq. Detection is sufficient for cataloging sites; quantification is necessary for comparing modification levels between conditions.

2. When is m6A detection sufficient for my project?

Detection is sufficient when the primary goal is to identify where m6A sites are located — cataloging the m6A landscape in a new organism, tissue, or cell type. It is also appropriate when RNA input is severely limited (nanogram quantities), when long-read isoform analysis is the priority, or when screening multiple conditions for large-effect differences before committing to quantitative follow-up.

3. How does GLORI-seq achieve single-base m6A quantification?

GLORI-seq uses glyoxal and nitrite to chemically deaminate unmethylated adenosines to inosines (read as guanosine in sequencing). m6A residues are protected from deamination by the methyl group and remain read as adenosine. The fraction of reads retaining A at each genomic position directly reports the proportion of transcripts carrying m6A at that site, providing absolute stoichiometric measurement at single-base resolution.

4. Can I compare m6A levels between conditions using detection methods?

Not reliably. Detection methods conflate changes in m6A modification with changes in transcript expression. A site detected in treatment but not in control could reflect a genuine increase in m6A, higher expression of the transcript, or a technical difference in IP efficiency. Quantitative methods measure modification fractions independently of expression, enabling direct comparison between conditions.

5. Is single-base m6A quantification more expensive than detection?

Yes, quantitative methods generally require deeper sequencing and more complex library preparation, resulting in higher per-sample costs. However, the per-site information content is substantially higher — a quantitative dataset directly supports mechanistic inference, while a detection dataset may require additional experiments to interpret. For projects focused on condition comparison, the quantitative approach is often more cost-effective when downstream validation costs are considered.

References

1. Liu, Cong, Hanxiao Sun, Yunpeng Yi, Weiguo Shen, Kai Li, Ye Xiao, Fei Li, et al. "Absolute quantification of single-base m6A methylation in the mammalian transcriptome using GLORI." Nature Biotechnology, vol. 41, 2023, pp. 355–366. DOI: 10.1038/s41587-022-01487-9
2. Garcia-Campos, Miguel Angel, Sarit Edelheit, Ursula Toth, Modi Safra, Ran Shachar, Sergey Viukov, Roni Winkler, et al. "Deciphering the 'm6A Code' via Antibody-Independent Quantitative Profiling." Cell, vol. 178, no. 3, 2019, pp. 731–747.e16. DOI: 10.1016/j.cell.2019.06.013
3. Meyer, Kate D. "DART-seq: an antibody-free method for global m6A detection." Nature Methods, vol. 16, no. 12, 2019, pp. 1275–1280. DOI: 10.1038/s41592-019-0570-0
4. Linder, Bastian, Anya V. Grozhik, Anthony O. Olarerin-George, Cem Meydan, Christopher E. Mason, and Samie R. Jaffrey. "Single-nucleotide-resolution mapping of m6A and m6Am throughout the transcriptome." Nature Methods, vol. 12, no. 8, 2015, pp. 767–772. DOI: 10.1038/nmeth.3453

CD Genomics provides epitranscriptomics services for research use only and they are not intended for clinical diagnosis or treatment.