Why modification-aware RNA sequencing is essential for safety and compliance.
Regulatory scrutiny of the integrity of RNA therapeutics is expected to increase and the regulatory framework for this is in the process of being defined. Modified bases in small RNAs and RNA therapeutics are crucial for processing and biological activity through complex feedback networks involving their writing, erasing, and sensing—disruption of which results in cellular dysregulation and disease.
For small-RNA drugs the pattern and quality of intended modifications is foundational to their effectiveness and toxicity. Modified mRNA bases are also used by cells for the regulation of half-life. Some mRNA therapies like vaccines are intentionally engineered with modified bases at specific positions to ensure persistence, while others are not; in either case, preparations can include RNAs with unintentional base modifications that create adverse events.
Authorities are signaling interest in standardized approaches to ensure the safety and efficacy of RNA therapeutics including the detection and quantification of both intended and unintended base modifications, prompting cals for more rigorous, modificationaware analyticalassays in submission. The problem is that while common contaminants such as dsRNA or plasmid DNA associated with the manufacturing process are easy to detect with existing technologies (e.g. LC-MS, PCRor chromatography), the discernment and quantification of modified nucleotides requires direct, de-novo RNA sequencing which has not yet been possible. This is the problem we elegantly solve at Directseq.
Micro-ribonucleotides (miRNAs) regulate cellular homeostasis primarily through gene silencing. Modified ribonucleotide bases (e.g., m6A) are critical for this function, forming binding conformations for proteins involved in writing, erasing, and sensing. Because these pathways are tightly controlled in cells, quality assurance of intended base-modification distributions in small-RNA therapeutics is vital for patient safety.
Both phosphoramidite syntheses and IVT reactions can produce trace isomer contaminants containing unnatural ribonucleotide modifications via chemical degradation (oxidation, deamination, hydrolysis), supplier or lot impurities, and polymerase side-activities. Lot-to-lot variation in these isomers has been associated with serious adverse events—including, in rare cases, death.
Common contaminants like dsRNA or plasmid DNA are detectable by LC-MS, PCR, or chromatography, but modified nucleotide discernment and quantification requires direct, de-novo RNA sequencing— which has not been possible with conventional methods.
Authorities are signaling interest in standardized, modification-aware assays in submissions to ensure the safety and efficacy of RNA therapeutics. Lack of visibility into intended and unintended base modifications creates critical risk and uncertainty for developers and regulators.
Liquid Chromatography Mass Spec(LC-MS)is used for controlof amajority of the RNA therapeutic CriticalQuality Attributes (CQAs), but sequence identity assessments are normaly performed through RT-PCR, which is incapable of identifying nucleotide base modifications and quantifying the isomers containing them. Modified base isomer distributions should be added to the CriticalQuality Attribute list for RNA therapeutics, and these distributions can only be appreciated through direct RNA sequencing. The problem with direct RNA sequencing up until now isthat sequence could only be obtained from highly purified samples capable of producing complete ladders, and only in aconfirmatory manner which required pre-determined knowledge of the sequence. Incapable of simultaneous, direct, de novo (and hypothesis-free)sequencing and quantitative assessment, prior state-of the-art platforms for RNA sequencing such as LC-MSare incapable of rigorous quality control needed to identify unknown contaminating modified base. The RNA therapeutics Industry looks forward to the introduction of 'novelanalytical tools to gain more insight into the use of CQAs (CriticalQuality Attributes) such as sequence identity' . The most attractive tools wilalow addressing large numbers of CQAssimultaneously, such as LC-MScurrently does, while also opening doors beyond current limitations to address previously unaddressed facets of CQAs (such as modified base containing isomers)that have emerged during clinicaland post clinical stages as potential safety concerns.
DirectSeq's Next Generation Mass Spec (DSMX™-Seq)platform is the significant technology leap patients, regulators and RNA therapeutics professionals have been waiting for. Instead of inferring RNA identity through mass readout, as with LC-MS, DSMX™-Seq-Seq reveals identity directly via de-novo sequencing. DSMX™-Seq-Seq delivers single-run, exhaustive, direct, 100% accurate sequencing of al RNA species in asample, including those containing any of up to 170+ base modifications. DSMX™-Seq-Seq alows for sequencing and quantification of not just the major RNA species, but each minor primary sequence isomer, including those containing one ribonucleotide base modification, multiple (even clustered) base modifications, and it delivers this sequence information while also delivering everything else that LC-MS does in terms of detection readout (e.g. presence of sequencing length, secondary and tertiary structure variants, cofactors etc.). This represents asignificant advancement beyond the limitations of LC-MS, and opens the door for the first time to empirical, hypothesis- and bias-free assessment of RNA identity and quality. There is no modified base information loss due to a cDNA conversion step, as with RNAseq, and unlike with pore based technologies, modified ribonucleotide bases are sequenced whether they are isolated on the RNA strand or clustered.
In 2024, we introduced DSMX™-Seq-Seq method asthe first solution to this problem. The 2D DSMX™-Seq-Seq platform overcame limitations associated the perfect ladder requirement, which only dominant RNA species can typicaly meet, and for the first time, extended sequencing reach of RNA therapeutic preparations to minor RNA isomerscontaining any number of canonicalor modified ribonucleotide base present in the sample. One of the first projects we worked on showed that even 'pure' tRNA preparations purchased from prominent vendors were unexpectedly mixed with minor isoforms containing modified ribonucleotide bases. In 2025, we extended the power of 2D DSMX™-Seq-Seq to three dimensionsfor more powerfulde-novo sequence detection of minor RNA species. Our new 3D DSMX™-Seq platform is now capable of not just detecting but sequencing and quantifiably (stoichiometricaly) mapping every RNA sequence in atherapeutic preparation, including minor modified base containing isomers, from more complex celular or IVTRNA preparations.
DSMX™-Seq generates complete sequence for species down to < 1% abundance and some of the very first applications of DSMX™-Seq showed that even pure tRNA and sgRNA samplespurchased from reputable vendors(including those synthesized with phorphoramidite chemistry) contain modified base isomers at levels far greater than 1% . With obvious implications for elevating the state of the science for RNA therapeutics quality control, we intend to establish DSMX™-Seq as astandard-bearer for reducing lot to lot variability in contaminating RNA species that contribute to unintended off target effects.
3D DSMXTM-Seq is the first RNA sequencing technology capable of directly sequencing and quantifying every sequence of amixed RNA sample, including minor isomers (down to 1%)differing by only asingle canonicalor modified base. 3D DSMXTM-Seq can identify alof the 170 known base modifications, whether only one or multiple are present, and unlike nanopore technologies, can read modified base sequences found in clusters, as they are commonly found in nature . It is adirect sequencing technology (no cDNA step)and unlike other RNA sequencing technologies, the accuracy has been demonstrated at 100% for smalmodified nucleotide reference libraries.
The RNA ishydrolyzed into ladder fragments, which are resolved for the various isoforms by mass, intensity and retention time. Within each layer, short sequencing reads are generated de novo by base-caling nucleotides from mass differences between adjacent ladder fragments. The short reads are then assembled into ful length RNA sequences. We routinely sequence siRNA, miRNA, CRISPR/Cas9 sgRNAs and mRNAswhether they contain modified ribonucleotide bases or not. We can also distinguish native from smal-molecular or other species-bound RNAs for RNA targeted drug developers.
Recent projects have sequenced tRNA, synthetic oligoribonucleotides, siRNA/miRNA mimics, sgRNA and tsRNA samples, and we are completing the validation work necessary to next take on mRNA samplesin the near future. Our tsRNA work provided crucialgranulation for noteworthy research published in Nature Communications, demonstrating the disruptive and transformative nature of our technology.
One of the first demonstrations of DSMX™-Seq showcased its discriminatory sequencing power. We found that, surprisingly, acommercialy provided research grade tRNA sample from areputable vendor wasactualy aheterogeneous mixture of five different length, canonicalribonucleotide base, and modified ribonucleotide base isomers. One of the two canonicalbase substitution isomers was present at a relative abundance of 0.6 (the other at 0.1), and two different modified base isomers were present at arelative abundance of about 0.2 . Later work demonstrated the homogeneouspurity of asynthetic 20bp ribonucleotide oligomer, its sensitivity and accuracy confirming not just sequence but stoichiometry of synthetic ribonucleotide oligomer mixtures.
One of the most important demonstrations of DSMX™-Seq showed that a100 nucleotide modified base sgRNA sample produced using phosphoramidite chemical synthesis was revealed to be amixture of the intended, major species along with impurity isomers arising from phosphoramidite chemistry. One length isomer (a truncation) represented 20.7% of the sample mass, and three unintended modified ribonucleotide containing isomers were identified to represent 5.0%, 4.5% and 0.3% of the sample. This result is meaningfulto the RNA therapeutics industry, where ribonucleotide base integrity and purity of RNA drugs is generaly assumed when synthesized chemicaly, and suggests that such assumptions need to be revisited. Wehave also demonstrated sequencing on endogenous tsRNAs isolated from the liver of mice. A tsRNA-Glue-CTC species was isolated and shown to be present in four distinct modified ribonucleotide base isoforms, each containing asingle mC, D or D modified basesat positions 6, 19 and 20, respectively.
The next generation of mRNA drugsare already expanding beyond vaccines to additional disease states, staging the power of this revolutionary technology to fi l critical therapeutic gaps. But adverse events linked with particular delivery modalities and RNA contaminants are just now being fuly understood, and a regulatory framework is stilbeing defined. The major contaminants length truncations, dsRNA and plasmid template carryover from manufacturing are relatively easy to detect, but the same is not the case with modified nucleotide isomers. It has long been known that even low levelsof RNA therapeutic secondary and tertiary sequence isomerization can lead can lead to innate immune sensing, but mechanistic studies show that aberrantly modified ribonucleotides can lead to reduced regulatory control of translation efficiency, altered degradation kinetics, and even nuclear translocation and DNA integration , producing outsized biological effects relative to their abundance.
The emergence of unusualy high adverse event rates for COVID vaccineswas a wake-up caland heightened public concern and scrutiny of RNA therapeutics. The COVID-19 mRNA vaccinescontained intentionaly engineered ribonucleotide modifications to enhance durability and reduce unintended immune response, but showed unexpectedly long spike protein mRNA persistence up to severalmonths far longer than intended and lot to lot variation in contamination has been associated with varying levels of serious adverse events . Similar observations have been made for monogenicreplacement RNA therapies.
Product-specific impurities with unnaturalhalf-lives are a concern because their presences is not easily contro lable or knowable with current analytic instrumentation, and their effects not entirely testable in clinical trials of limited patient populations. Modified nucleotide incorporation in RNA therapeutics therefore merit carefulanalytical control and correlation with clinical safety data and we can expect an enhanced Quality Assurance framework for detecting and preventing their contribution to RNA therapeutics contamination going forward. This wilrequire new standards be developed, and the adoption of new technologies and proceduresto enable more comprehensive assessments of RNA complexity during lot analysis, including:
RNA molecules encode biological information not only through their sequences but also through over 170 known chemical modifications1. These modifications influence RNA structure, stability, translation, and function—and are implicated in over 100 human diseases, including cancer, diabetes, Alzheimer's, and Parkinson's. As such, RNA and its modification profiles are emerging as powerful biomarkers and therapeutic targets.
Yet, our understanding of RNA sequence and modification diversity remains limited2. Current sequencing technologies offer partial insights but fail to provide the fu l spectrum of RNA sequence variants. In fact, we do not know how many unique RNA molecules or sequence variants are present in a sample, and further, we do not know the complete sequence content of each RNA, including the identity and location of every nucleotide (canonical or modified) within a full-length RNA.
To overcome these limitations, we developed NGMS-Seq, anext-generation mass spectrometry-based sequencing platform. Unlike optical or electronic systems such as Illumina or Nanopore-based RNA sequencing, NGMS-Seq uses mass spectrometry as direct readout to comprehensively sequence full-length RNAs and their modifications—without bias or prior knowledge across isolated and bulk RNA samples.
These capabilities pave the way for the world's first Human RNome Project, alandmark initiative to draft the first complete sequence of all human RNA molecules and their modifications.Just as Sanger sequencing enabled the Human Genome Project (completed in 2003), NGMS-Seq has potential to deliver the first complete Human RNome, transforming RNA biology, therapeutics, and diagnostics.