The RNA inside a cell carries a second layer of information. More than 170 different chemical modifications decorate its building blocks, editing how RNA folds, how efficiently it is translated, how quickly it degrades, and how cells respond to stress. This so-called epitranscriptome is implicated in neurological disorders, cancer, immune disease, and drug resistance, yet it remains poorly mapped, largely because the instruments used to study it can see modifications but cannot count them. A team led by MIT’s Peter Dedon, with co-first authors Junzhou Wu and Jingjing Sun, has now adapted a workhorse quantification technology from proteomics to RNA, enabling precise, site-resolved, multiplexed measurement of RNA modifications. The method, described in Nature Communications, quantified 22 modification sites in bacterial transfer RNA with coefficients of variation below 5 percent, including two sites never mapped before.
The technical problem is older than the field’s name. Most existing methods for locating RNA modifications rely on indirect signals. Antibody-based enrichment, such as the widely used m6A-seq, pulls down modified fragments but reads them through reverse transcription, whose misincorporation or premature stops are the actual signal. Nanopore sequencing infers modifications from ionic current changes as RNA threads through a pore. Both approaches identify edits by proxy, which makes them powerful for discovery but weak for quantification: comparing modification levels across conditions, treatments, or genetic backgrounds has been notoriously unreliable. Label-free mass spectrometry measures modification mass directly but suffers run-to-run variability, and isotopic labeling approaches have struggled with retention-time shifts and poor multiplexing.
The new platform, which the authors call RMT, for RNA-specific isobaric tandem mass tagging, borrows the logic of the TMT and iTRAQ reagents that transformed quantitative proteomics. Each digested RNA sample is labeled with a unique tag. All the tags have identical total mass, so they are isobaric, but each carries a chemically distinct reporter group. Multiplexed samples co-elute from the chromatograph and appear as a single peak in the first mass spectrum; when the instrument fragments the molecules, it releases the reporter ions, whose intensities reveal how much of each modification came from each sample. Ten or more conditions can be compared in a single run, eliminating the run-to-run noise that plagues label-free analysis.
Adapting the tags to RNA required solving a chemistry problem the proteomics world never faced. The researchers attached the isobaric tags to a pCp scaffold, the standard 3-prime-end labeling reagent, and ligated them onto RNA fragments enzymatically with T4 RNA ligase. The approach is dramatically cheaper than the chemical labeling used in proteomics, roughly 10 nanomoles of reagent per reaction versus about 2 micromoles, and it leaves the RNA’s modifications untouched. Identification comes from nucleobase fragment analysis, which reads modification identity directly from molecular weight, and the entire data pipeline runs on open-source software with commercially available reagents.
The proof-of-concept used transfer RNA from the bacterium Pseudomonas aeruginosa strain PA14, a model organism with a rich modification landscape. The work was carried out across MIT’s Department of Biological Engineering, the Singapore-MIT Alliance for Research and Technology, and the University of Florida. The platform quantified 22 modification sites across 65 of the 66 predicted tRNA genes, with coefficients of variation below 5 percent, and it distinguished closely related modifications that older methods conflate, such as isopentenyl-adenosine and its hydroxylated form. Among the newly characterized sites were m2A38 and Gm/Cm39, previously undescribed positions whose writer enzymes the team then identified by screening knockout strains, assigning one writer, TrmV, to m2A38 and another, TrmS, to the methylations at position 39. Knockout experiments also revealed interdependence between modifications: deleting one writer enzyme shifted patterns at other sites, and shifting cells between rich and minimal growth media produced dynamic modification changes consistent with a role in stress adaptation.
The epitranscriptome becomes a measurable signal rather than a static catalog. With multiplexed quantification, researchers can ask how modification levels change across time, conditions, and genotypes, the kinds of experiments that turned proteomics into a quantitative science. The authors are candid about the method’s limits: reporter-ion crosstalk and co-isolation interference, well characterized in proteomics, constrain accuracy and dynamic range; the TMT-derived tags show an isotope-based retention-time shift on RNA that must be engineered around; and validation so far is limited to tRNA. The manuscript is an early-access version, published before final copyediting.
The work transfers analytical maturity from proteins to RNA. Proteomics spent two decades learning to count, and its isobaric tagging chemistry is now standard in thousands of labs. Carrying that same chemistry across to the epitranscriptome, at a fraction of the reagent cost, gives biologists a way to treat RNA modifications as quantitative biology, not just a list of where they sit. The second genetic code, it turns out, is finally becoming readable in numbers.
Sources: Wu, J., Sun, J., Yuan, Y. et al. Quantitative RNA modification mapping by mass spectrometry with isobaric tags and nucleobase fragment analysis. Nature Communications (2026). DOI: 10.1038/s41467-026-76537-w.