Distinguishing RNA modifications from noise in epitranscriptome maps

Current progress in strategies to profile transcriptomic m6A modifications

Various methods have been developed so far for detecting N 6-methyladenosine (m6A). The total m6A level or the m6A status at individual positions on mRNA can be detected and quantified through some sequencing-independent biochemical methods, such as LC/MS, SCARLET, SELECT, and m6A-ELISA. However, the m6A-detection techniques relying on high-throughput sequencing have more effectively advanced the understanding about biological significance of m6A-containing mRNA and m6A pathway at a transcriptomic level over the past decade. Various SGS-based (Second Generation Sequencing-based) methods with different detection principles have been widely employed for this purpose. These principles include m6A-enrichment using antibodies, discrimination of m6A from unmodified A-base by nucleases, a fusion protein strategy relying on RNA-editing enzymes, and marking m6A with chemical/biochemical reactions. Recently, TGS-based (Third Generation Sequencing-based) methods have brought a new trend by direct m6A-detection. This review first gives a brief introduction of current knowledge about m6A biogenesis and function, and then comprehensively describes m6A-profiling strategies including their principles, procedures, and features. This will guide users to pick appropriate methods according to research goals, give insights for developing novel techniques in varying areas, and continue to expand our boundary of knowledge on m6A.

Sequencing accuracy and systematic errors of nanopore direct RNA sequencing

BACKGROUND

Direct RNA sequencing (dRNA-seq) on the Oxford Nanopore Technologies (ONT) platforms can produce reads covering up to full-length gene transcripts, while containing decipherable information about RNA base modifications and poly-A tail lengths. Although many published studies have been expanding the potential of dRNA-seq, its sequencing accuracy and error patterns remain understudied.

RESULTS

We present the first comprehensive evaluation of sequencing accuracy and characterisation of systematic errors in dRNA-seq data from diverse organisms and synthetic in vitro transcribed RNAs. We found that for sequencing kits SQK-RNA001 and SQK-RNA002, the median read accuracy ranged from 87% to 92% across species, and deletions significantly outnumbered mismatches and insertions. Due to their high abundance in the transcriptome, heteropolymers and short homopolymers were the major contributors to the overall sequencing errors. We also observed systematic biases across all species at the levels of single nucleotides and motifs. In general, cytosine/uracil-rich regions were more likely to be erroneous than guanines and adenines. By examining raw signal data, we identified the underlying signal-level features potentially associated with the error patterns and their dependency on sequence contexts. While read quality scores can be used to approximate error rates at base and read levels, failure to detect DNA adapters may be a source of errors and data loss. By comparing distinct basecallers, we reason that some sequencing errors are attributable to signal insufficiency rather than algorithmic (basecalling) artefacts. Lastly, we generated dRNA-seq data using the latest SQK-RNA004 sequencing kit released at the end of 2023 and found that although the overall read accuracy increased, the systematic errors remain largely identical compared to the previous kits.

CONCLUSIONS

As the first systematic investigation of dRNA-seq errors, this study offers a comprehensive overview of reproducible error patterns across diverse datasets, identifies potential signal-level insufficiency, and lays the foundation for error correction methods.

m6ACali: machine learning-powered calibration for accurate m6A detection in MeRIP-Seq

We present m6ACali, a novel machine-learning framework aimed at enhancing the accuracy of N6-methyladenosine (m6A) epitranscriptome profiling by reducing the impact of non-specific antibody enrichment in MeRIP-Seq. The calibration model serves as a genomic feature-based classifier that refines the identification of m6A sites, distinguishing those genuinely present from those that can be detected in in-vitro transcribed (IVT) control experiments. We find that m6ACali effectively identifies non-specific binding peaks reported by exomePeak2 and MACS2 in novel MeRIP-Seq datasets without the need for paired IVT controls. The model interpretation revealed that off-target antibody binding sites commonly occur at short exons and short mRNAs, originating from high read coverage regions that share the motif sequence with true m6A sites. We also reveal that the ML strategy can efficiently adjust differentially methylated peaks and other antibody-dependent, base-resolution m6A detection techniques. As a result, m6ACali offers a promising method for the universal enhancement of m6A profiles generated by MeRIP-Seq experiments, elevating the benchmark for omics-level m6A data integration.

2'-O-methylation (Nm) in RNA: progress, challenges, and future directions

RNA 2'-O-methylation (Nm) is highly abundant in noncoding RNAs including ribosomal RNA (rRNA), transfer RNA (tRNA), and small nuclear RNA (snRNA), and occurs in the 5' cap of virtually all messenger RNAs (mRNAs) in higher eukaryotes. More recently, Nm has also been reported to occur at internal sites in mRNA. High-throughput methods have been developed for the transcriptome-wide detection of Nm. However, these methods have mostly been applied to abundant RNAs such as rRNA, and the validity of the internal mRNA Nm sites detected with these approaches remains controversial. Nonetheless, Nm in both coding and noncoding RNAs has been demonstrated to impact cellular processes, including translation and splicing. In addition, Nm modifications at the 5' cap and possibly at internal sites in mRNA serve to prevent the binding of nucleic acid sensors, thus preventing the activation of the innate immune response by self-mRNAs. Finally, Nm has been implicated in a variety of diseases including cancer, cardiovascular diseases, and neurologic syndromes. In this review, we discuss current challenges in determining the distribution, regulation, function, and disease relevance of Nm, as well as potential future directions for the field.

SCARPET: site-specific quantification of methylated and nonmethylated adenosines reveals m6A stoichiometry

m6A has different stoichiometry at different positions in different mRNAs. However, the exact stoichiometry of m6A is difficult to measure. Here, we describe SCARPET (site-specific cleavage and radioactive-labeling followed by purification, exonuclease digestion, and thin-layer chromatography), a simple and streamlined biochemical assay for quantifying m6A at any specific site in any mRNA. SCARPET involves a site-specific cleavage of mRNA immediately 5' of an adenosine site in an mRNA. This site is radiolabeled with 32P, and after a series of steps to purify the RNA and to remove nonspecific signals, the nucleotide is resolved by TLC to visualize A and m6A at this site. Quantification of these spots reveals the m6A stoichiometry at the site of interest. SCARPET can be applied to poly(A)-enriched RNA, or preferably purified mRNA, which produces more accurate m6A stoichiometry measurements. We show that sample processing steps of SCARPET can be performed in a single day, and results in a specific and accurate measurement of m6A stoichiometry at specific sites in mRNA. Using SCARPET, we measure exact m6A stoichiometries in specific mRNAs and show that Zika genomic RNA lacks m6A at previously mapped sites. SCARPET will be useful for testing specific sites for their m6A stoichiometry and to assess how m6A stoichiometry changes in different conditions and cellular contexts.

MePMe-seq: antibody-free simultaneous m6A and m5C mapping in mRNA by metabolic propargyl labeling and sequencing

Internal modifications of mRNA have emerged as widespread and versatile regulatory mechanism to control gene expression at the post-transcriptional level. Most of these modifications are methyl groups, making S-adenosyl-L-methionine (SAM) a central metabolic hub. Here we show that metabolic labeling with a clickable metabolic precursor of SAM, propargyl-selenohomocysteine (PSH), enables detection and identification of various methylation sites. Propargylated A, C, and G nucleosides form at detectable amounts via intracellular generation of the corresponding SAM analogue. Integration into next generation sequencing enables mapping of N6-methyladenosine (m6A) and 5-methylcytidine (m5C) sites in mRNA with single nucleotide precision (MePMe-seq). Analysis of the termination profiles can be used to distinguish m6A from 2'-O-methyladenosine (Am) and N1-methyladenosine (m1A) sites. MePMe-seq overcomes the problems of antibodies for enrichment and sequence-motifs for evaluation, which was limiting previous methodologies. Metabolic labeling via clickable SAM facilitates the joint evaluation of methylation sites in RNA and potentially DNA and proteins.

The CTBP2-PCIF1 complex regulates m6Am modification of mRNA in head and neck squamous cell carcinoma

PCIF1 can mediate the methylation of N6,2'-O-dimethyladenosine (m6Am) in mRNA. Yet, the detailed interplay between PCIF1 and the potential cofactors and its pathological significance remain elusive. Here, we demonstrated that PCIF1-mediated cap mRNA m6Am modification promoted head and neck squamous cell carcinoma progression both in vitro and in vivo. CTBP2 was identified as a cofactor of PCIF1 to catalyze m6Am deposition on mRNA. CLIP-Seq data demonstrated that CTBP2 bound to similar mRNAs as compared with PCIF1. We then used the m6Am-Seq method to profile the mRNA m6Am site at single-base resolution and found that mRNA of TET2, a well-known tumor suppressor, was a major target substrate of the PCIF1-CTBP2 complex. Mechanistically, knockout of CTBP2 reduced PCIF1 occupancy on TET2 mRNA, and the PCIF1-CTBP2 complex negatively regulated the translation of TET2 mRNA. Collectively, our study demonstrates the oncogenic function of the epitranscriptome regulator PCIF1-CTBP2 complex, highlighting the importance of the m6Am modification in tumor progression.

Epitranscriptomic subtyping, visualization, and denoising by global motif visualization

Advances in sequencing technologies have empowered epitranscriptomic profiling at the single-base resolution. Putative RNA modification sites identified from a single high-throughput experiment may contain one type of modification deposited by different writers or different types of modifications, along with false positive results because of the challenge of distinguishing signals from noise. However, current tools are insufficient for subtyping, visualization, and denoising these signals. Here, we present iMVP, which is an interactive framework for epitranscriptomic analysis with a nonlinear dimension reduction technique and density-based partition. As exemplified by the analysis of mRNA m5C and ModTect variant data, we show that iMVP allows the identification of previously unknown RNA modification motifs and writers and the discovery of false positives that are undetectable by traditional methods. Using putative m6A/m6Am sites called from 8 profiling approaches, we illustrate that iMVP enables comprehensive comparison of different approaches and advances our understanding of the difference and pattern of true positives and artifacts in these methods. Finally, we demonstrate the ability of iMVP to analyze an extremely large human A-to-I editing dataset that was previously unmanageable. Our work provides a general framework for the visualization and interpretation of epitranscriptomic data.

Navigating the pitfalls of mapping DNA and RNA modifications

Chemical modifications to nucleic acids occur across the kingdoms of life and carry important regulatory information. Reliable high-resolution mapping of these modifications is the foundation of functional and mechanistic studies, and recent methodological advances based on next-generation sequencing and long-read sequencing platforms are critical to achieving this aim. However, mapping technologies may have limitations that sometimes lead to inconsistent results. Some of these limitations are technical in nature and specific to certain types of technology. Here, however, we focus on common (yet not always widely recognized) pitfalls that are shared among frequently used mapping technologies and discuss strategies to help technology developers and users mitigate their effects. Although the emphasis is primarily on DNA modifications, RNA modifications are also discussed.

Methylated guanosine and uridine modifications in S. cerevisiae mRNAs modulate translation elongation

Chemical modifications to protein encoding messenger RNAs (mRNAs) influence their localization, translation, and stability within cells. Over 15 different types of mRNA modifications have been observed by sequencing and liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) approaches. While LC-MS/MS is arguably the most essential tool available for studying analogous protein post-translational modifications, the high-throughput discovery and quantitative characterization of mRNA modifications by LC-MS/MS has been hampered by the difficulty of obtaining sufficient quantities of pure mRNA and limited sensitivities for modified nucleosides. We have overcome these challenges by improving the mRNA purification and LC-MS/MS pipelines. The methodologies we developed result in no detectable non-coding RNA modifications signals in our purified mRNA samples, quantify 50 ribonucleosides in a single analysis, and provide the lowest limit of detection reported for ribonucleoside modification LC-MS/MS analyses. These advancements enabled the detection and quantification of 13 S. cerevisiae mRNA ribonucleoside modifications and reveal the presence of four new S. cerevisiae mRNA modifications at low to moderate levels (1-methyguanosine, N2-methylguanosine, N2,N2-dimethylguanosine, and 5-methyluridine). We identified four enzymes that incorporate these modifications into S. cerevisiae mRNAs (Trm10, Trm11, Trm1, and Trm2, respectively), though our results suggest that guanosine and uridine nucleobases are also non-enzymatically methylated at low levels. Regardless of whether they are incorporated in a programmed manner or as the result of RNA damage, we reasoned that the ribosome will encounter the modifications that we detect in cells. To evaluate this possibility, we used a reconstituted translation system to investigate the consequences of modifications on translation elongation. Our findings demonstrate that the introduction of 1-methyguanosine, N2-methylguanosine and 5-methyluridine into mRNA codons impedes amino acid addition in a position dependent manner. This work expands the repertoire of nucleoside modifications that the ribosome must decode in S. cerevisiae. Additionally, it highlights the challenge of predicting the effect of discrete modified mRNA sites on translation de novo because individual modifications influence translation differently depending on mRNA sequence context.

Critical functions of N6-adenosine methylation of mRNAs in T cells

The existence of N⁶-adenosine methylation (m⁶A) of mRNA has been known for a long time, but only recently its regulatory potential was uncovered. Current research deciphers the molecular determinants leading to the deposition of this modification and consequences for modified mRNAs. It also evaluates the importance of such modifications for specific cell types and programs. In this review, we summarize the current knowledge on m⁶A modification of mRNAs in conventional and regulatory T cells and T-cell-driven immune responses and pathology. We discuss the impact of m⁶A modification on T cell activation including cytokine and antigen receptor signaling or sensing of double-stranded RNAs (dsRNA).

Developmental mRNA m5C landscape and regulatory innovations of massive m5C modification of maternal mRNAs in animals

m⁵C is one of the longest-known RNA modifications, however, its developmental dynamics, functions, and evolution in mRNAs remain largely unknown. Here, we generate quantitative mRNA m⁵C maps at different stages of development in 6 vertebrate and invertebrate species and find convergent and unexpected massive methylation of maternal mRNAs mediated by NSUN2 and NSUN6. Using Drosophila as a model, we reveal that embryos lacking maternal mRNA m⁵C undergo cell cycle delays and fail to timely initiate maternal-to-zygotic transition, implying the functional importance of maternal mRNA m⁵C. From invertebrates to the lineage leading to humans, two waves of m⁵C regulatory innovations are observed: higher animals gain cis-directed NSUN2-mediated m⁵C sites at the 5' end of the mRNAs, accompanied by the emergence of more structured 5'UTR regions; humans gain thousands of trans-directed NSUN6-mediated m⁵C sites enriched in genes regulating the mitotic cell cycle. Collectively, our studies highlight the existence and regulatory innovations of a mechanism of early embryonic development and provide key resources for elucidating the role of mRNA m⁵C in biology and disease.

mRNAs are known to be decorated with m5C at a low-to-medium level. Here, the authors generate atlases of mRNA m5C during animal development in 6 species and identify convergent and unexpected massive methylation of maternal mRNAs by NSUN2 and NSUN6.

The Non-N6-Methyladenosine Epitranscriptome Patterns and Characteristics of Tumor Microenvironment Infiltration and Mesenchymal Transition in Glioblastoma

Background

An increasing number of RNA modification types other than N⁶-methyladenosine (m⁶A) modification have been detected. Nonetheless, the probable functions of RNA modifications beyond m⁶A in the tumor microenvironment (TME), mesenchymal (MES) transition, immunotherapy, and drug sensitivity remain unclear.

Methods

We analyzed the characteristics of 32 non-m⁶A RNA modification regulators in 539 glioblastoma (GBM) patients and the TME cell infiltration and MES transition patterns. Using principal component analysis, a non-m⁶A epitranscriptome regulator score (RM score) model was established. We estimated the association between RM score and clinical characteristics, TME status, GBM subtypes, and drug and immunotherapy response.

Results

Three definite non-m⁶A RNA modification patterns associated with diverse biological pathways and clinical characteristics were identified. The high RM score group was characterized by a poor prognosis, enhanced immune infiltration, and MES subtype. Further analysis indicated that the high RM score group had a lower tumor mutation burden as well as a weaker response to immunotherapy. The higher RM score group may benefit more from drugs targeting the EGFR and WNT signaling pathways.

Conclusion

Our study exposed the potential relationship of non-m⁶A RNA modification regulators with clinical features, TME status, and GBM subtype and clarified its therapeutic value.

Control of animal virus replication by RNA adenosine methylation

Methylation at the N⁶-position of either adenosine (m⁶A) or 2'-O-methyladenosine (m⁶Am) represents two of the most abundant internal modifications of coding and non-coding RNAs, influencing their maturation, stability and function. Additionally, although less abundant and less well-studied, monomethylation at the N¹-position (m¹A) can have profound effects on RNA folding. It has been known for several decades that RNAs produced by both DNA and RNA viruses can be m⁶A/m⁶Am modified and the list continues to broaden through advances in detection technologies and identification of the relevant methyltransferases. Recent studies have uncovered varied mechanisms used by viruses to manipulate the m⁶A pathway in particular, either to enhance virus replication or to antagonize host antiviral defenses. As such, RNA modifications represent an important frontier of exploration in the broader realm of virus-host interactions, and this new knowledge already suggests exciting opportunities for therapeutic intervention. In this review we summarize the principal mechanisms by which m⁶A/m⁶Am can promote or hinder viral replication, describe how the pathway is actively manipulated by biomedically important viruses, and highlight some remaining gaps in understanding how adenosine methylation of RNA controls viral replication and pathogenesis.

A bifunctional chemical signature enabling RNA 4-thiouridine enrichment sequencing with single-base resolution

Both sequence enrichment and base resolution are essential for accurate sequencing analysis of low-abundance RNA. Yet they are hindered by the lack of molecular tools. Here we report a bifunctional chemical signature for RNA 4-thiouridine (4sU) enrichment sequencing with single-base resolution. This chemical signature is designed for specific 4sU labeling with two functional parts. One part is a distal alkynyl group for the biotin-assisted pulldown enrichment of target molecules via click chemistry crosslinking. The other part is a -NH group proximal to the pyrimidine ring of 4sU. It allows 4sU-to-cytosine transition during the polymerase-catalyzed extension reaction based on altering hydrogen-bonding patterns. Ultimately, the 4sU-containing RNA molecules can be enriched and accurately analyzed by single-base resolution sequencing. The proposed method also holds great potential to investigate transcriptome dynamics integrated with high-throughput sequencing.

Transcription-wide mapping of dihydrouridine reveals that mRNA dihydrouridylation is required for meiotic chromosome segregation

The epitranscriptome has emerged as a new fundamental layer of control of gene expression. Nevertheless, the determination of the transcriptome-wide occupancy and function of RNA modifications remains challenging. Here we have developed Rho-seq, an integrated pipeline detecting a range of modifications through differential modification-dependent rhodamine labeling. Using Rho-seq, we confirm that the reduction of uridine to dihydrouridine (D) by the Dus reductase enzymes targets tRNAs in E. coli and fission yeast. We find that the D modification is also present on fission yeast mRNAs, particularly those encoding cytoskeleton-related proteins, which is supported by large-scale proteome analyses and ribosome profiling. We show that the α-tubulin encoding mRNA nda2 undergoes Dus3-dependent dihydrouridylation, which affects its translation. The absence of the modification on nda2 mRNA strongly impacts meiotic chromosome segregation, resulting in low gamete viability. Applying Rho-seq to human cells revealed that tubulin mRNA dihydrouridylation is evolutionarily conserved.

Detection of m6A from direct RNA sequencing using a multiple instance learning framework

RNA modifications such as m6A methylation form an additional layer of complexity in the transcriptome. Nanopore direct RNA sequencing can capture this information in the raw current signal for each RNA molecule, enabling the detection of RNA modifications using supervised machine learning. However, experimental approaches provide only site-level training data, whereas the modification status for each single RNA molecule is missing. Here we present m6Anet, a neural-network-based method that leverages the multiple instance learning framework to specifically handle missing read-level modification labels in site-level training data. m6Anet outperforms existing computational methods, shows similar accuracy as experimental approaches, and generalizes with high accuracy to different cell lines and species without retraining model parameters. In addition, we demonstrate that m6Anet captures the underlying read-level stoichiometry, which can be used to approximate differences in modification rates. Overall, m6Anet offers a tool to capture the transcriptome-wide identification and quantification of m6A from a single run of direct RNA sequencing.

mRNA, a Revolution in Biomedicine

The perspective of using messenger RNA (mRNA) as a therapeutic molecule first faced some uncertainties due to concerns about its instability and the feasibility of large-scale production. Today, given technological advances and deeper biomolecular knowledge, these issues have started to be addressed and some strategies are being exploited to overcome the limitations. Thus, the potential of mRNA has become increasingly recognized for the development of new innovative therapeutics, envisioning its application in immunotherapy, regenerative medicine, vaccination, and gene editing. Nonetheless, to fully potentiate mRNA therapeutic application, its efficient production, stabilization and delivery into the target cells are required. In recent years, intensive research has been carried out in this field in order to bring new and effective solutions towards the stabilization and delivery of mRNA. Presently, the therapeutic potential of mRNA is undoubtedly recognized, which was greatly reinforced by the results achieved in the battle against the COVID-19 pandemic, but there are still some issues that need to be improved, which are critically discussed in this review.

Metabolic turnover and dynamics of modified ribonucleosides by 13C labeling

Tandem mass spectrometry (MS/MS) is an accurate tool to assess modified ribonucleosides and their dynamics in mammalian cells. However, MS/MS quantification of lowly abundant modifications in non-ribosomal RNAs is unreliable, and the dynamic features of various modifications are poorly understood. Here, we developed a ¹³C labeling approach, called ¹³C-dynamods, to quantify the turnover of base modifications in newly transcribed RNA. This turnover-based approach helped to resolve mRNA from ncRNA modifications in purified RNA or free ribonucleoside samples and showed the distinct kinetics of the N6-methyladenosine (m⁶A) versus 7-methylguanosine (m⁷G) modification in polyA+-purified RNA. We uncovered that N6,N6-dimethyladenosine (m⁶₂A) exhibits distinct turnover in small RNAs and free ribonucleosides when compared to known m⁶₂A-modified large rRNAs. Finally, combined measurements of turnover and abundance of these modifications informed on the transcriptional versus posttranscriptional sensitivity of modified ncRNAs and mRNAs, respectively, to stress conditions. Thus, ¹³C-dynamods enables studies of the origin of modified RNAs at steady-state and subsequent dynamics under nonstationary conditions. These results open new directions to probe the presence and biological regulation of modifications in particular RNAs.

Systematic calibration of epitranscriptomic maps using a synthetic modification-free RNA library

Recent years have witnessed rapid progress in the field of epitranscriptomics. Functional interpretation of the epitranscriptome relies on sequencing technologies that determine the location and stoichiometry of various RNA modifications. However, contradictory results have been reported among studies, bringing the biological impacts of certain RNA modifications into doubt. Here, we develop a synthetic RNA library resembling the endogenous transcriptome but without any RNA modification. By incorporating this modification-free RNA library into established mapping techniques as a negative control, we reveal abundant false positives resulting from sequence bias or RNA structure. After calibration, precise and quantitative mapping expands the understanding of two representative modification types, N⁶-methyladenosine (m⁶A) and 5-methylcytosine (m⁵C). We propose that this approach provides a systematic solution for the calibration of various RNA-modification mappings and holds great promise in epitranscriptomic studies.

General Principles for the Detection of Modified Nucleotides in RNA by Specific Reagents

Epitranscriptomics heavily rely on chemical reagents for the detection, quantification, and localization of modified nucleotides in transcriptomes. Recent years have seen a surge in mapping methods that use innovative and rediscovered organic chemistry in high throughput approaches. While this has brought about a leap of progress in this young field, it has also become clear that the different chemistries feature variegated specificity and selectivity. The associated error rates, e.g., in terms of false positives and false negatives, are in large part inherent to the chemistry employed. This means that even assuming technically perfect execution, the interpretation of mapping results issuing from the application of such chemistries are limited by intrinsic features of chemical reactivity. An important but often ignored fact is that the huge stochiometric excess of unmodified over-modified nucleotides is not inert to any of the reagents employed. Consequently, any reaction aimed at chemical discrimination of modified versus unmodified nucleotides has optimal conditions for selectivity that are ultimately anchored in relative reaction rates, whose ratio imposes intrinsic limits to selectivity. Here chemical reactivities of canonical and modified ribonucleosides are revisited as a basis for an understanding of the limits of selectivity achievable with chemical methods.

METTL3-mediated m6A RNA methylation promotes the anti-tumour immunity of natural killer cells

Natural killer (NK) cells exert critical roles in anti-tumor immunity but how their functions are regulated by epitranscriptional modification (e.g., N6-methyladenosine (m6A) methylation) is unclear. Here we report decreased expression of the m6A "writer" METTL3 in tumor-infiltrating NK cells, and a positive correlation between protein expression levels of METTL3 and effector molecules in NK cells. Deletion of Mettl3 in NK cells alters the homeostasis of NK cells and inhibits NK cell infiltration and function in the tumor microenvironment, leading to accelerated tumor development and shortened survival in mice. The gene encoding SHP-2 is m6A modified, and its protein expression is decreased in METTL3-deficient NK cells. Reduced SHP-2 activity renders NK cells hyporesponsive to IL-15, which is associated with suppressed activation of the AKT and MAPK signaling pathway in METTL3-deficient NK cells. These findings show that m6A methylation safeguards the homeostasis and tumor immunosurveillance function of NK cells.

Analysis of pseudouridines and other RNA modifications using HydraPsiSeq protocol

Detection of RNA modified nucleotides using deep sequencing can be performed by several approaches, including antibody-driven enrichment and natural or chemically induced RT signatures. However, only very few RNA modified nucleotides generate natural RT signatures and antibody-driven enrichment heavily depends on the quality of antibodies used and may be highly biased. Thus, the use of chemically-induced RT signatures is now considered as the most trusted experimental approach. In addition, the use of chemical reagents allows inclusion of simple "mock-treated" controls, to exclude spontaneous RT arrests, SNPs and other misincorporation-prone sites. Hydrazine is a well-known RNA-specific reagent, already extensively used in the past for RNA sequencing and structural probing. Hydrazine is highly reactive to U and shows low reaction rates with ψ residues, allowing their distinction by deep sequencing-based protocols. However, other modified RNA residues also show particular behavior upon hydrazine treatment. Here we present methodological developments allowing to use HydraPsiSeq for precise quantification of RNA pseudouridylation and also detection and quantification of some other RNA modifications, in addition to ψ residues.

Arrow pushing in RNA modification sequencing

Methods to accurately determine the location and abundance of RNA modifications are critical to understanding their functional role. In this review, we describe recent efforts in which chemical reactivity and next-generation sequencing have been integrated to detect modified nucleotides in RNA. For eleven exemplary modifications, we detail chemical, enzymatic, and metabolic labeling protocols that can be used to differentiate them from canonical nucleobases. By emphasizing the molecular rationale underlying these detection methods, our survey highlights new opportunities for chemistry to define the role of RNA modifications in disease.

An N6-Methyladenosine-Related Gene Set Variation Score as a Prognostic Tool for Lung Adenocarcinoma

N6-methyladenosine (m6A) is the most prevalent type of RNA modification, and we hypothesized that patterns of m6A-related genes may be useful for estimating risk of lung adenocarcinoma (LUAD). An m6A-related gene set variation score (m6A-GSVS) was generated using RNA-sequencing data from LUAD patients in The Cancer Genome Atlas (TCGA). We investigated the association of m6A-GSVS with stemness, tumor mutational burden (TMB), expression of three immune checkpoints, levels of tumor-infiltrating lymphocytes (TILs), and patient prognosis. We found that m6A-GSVS was higher in LUAD than in healthy lung tissue, and it strongly correlated with stemness and TMB. Activated CD4 + T cells were more numerous in LUAD samples that had higher m6A-GSVS than in those with lower scores. Biological processes and pathways, including “Cell cycle,” “DNA replication,” and “RNA degradation,” were significantly enriched in samples with high scores. Furthermore, m6A-GSVS was an independent prognostic indicator in LUAD. In conclusion, we proposed an m6A-GSVS in LUAD. It is a putative indicator for evaluating the ability to RNA m6A, an independent prognostic indicator and associated with tumor stemness.

Micro-flow hydrophilic interaction liquid chromatography coupled with triple quadrupole mass spectrometry detects modified nucleosides in the transfer RNA pool of cyanobacteria

Post-transcriptional modification of nucleosides is observed in almost all elements of RNA. Modified nucleosides finely tune the structure of RNA molecules and affect vital functions, such as the modified wobble position 34 of transfer RNAs expanding the reading preference of anticodons to codons. Recent investigations have revealed that the modification species and their frequencies in an RNA element are not stable but vary with specific cellular factors including metabolites and particular proteins (writers, readers, and erasers). To understand the link between dynamic RNA modifications and biological processes, sensitive and reliable methods for determining modified nucleosides are required. In this study, micro-flow (8 μL/min) hydrophilic interaction liquid chromatography was coupled with triple quadrupole mass spectrometry for the simultaneous determination of adenosine, uridine, cytidine, guanosine, and 20 modified nucleosides. The method was calibrated using 0.1-1000 nM standards (∼0.03-300 ng/mL) and successfully applied to the determination of transfer RNA modifications in the model cyanobacterium Synechococcus elongatus PCC 7942. A protocol for the isolation of a clean transfer RNA pool was optimized, requiring only 25 ng for the identification and quantification of transfer RNA modifications. This micro-flow liquid chromatography-tandem mass spectrometry method constitutes the first step toward monitoring dynamic ribonucleoside modifications in a limited RNA sample.

METTL3 promotes tumour development by decreasing APC expression mediated by APC mRNA N6-methyladenosine-dependent YTHDF binding

The adenomatous polyposis coli (APC) is a frequently mutated tumour suppressor gene in cancers. However, whether APC is regulated at the epitranscriptomic level remains elusive. In this study, we analysed TCGA data and separated 200 paired oesophageal squamous cell carcinoma (ESCC) specimens and their adjacent normal tissues and demonstrated that methyltransferase-like 3 (METTL3) is highly expressed in tumour tissues. m⁶A-RNA immunoprecipitation sequencing revealed that METTL3 upregulates the m⁶A modification of APC, which recruits YTHDF for APC mRNA degradation. Reduced APC expression increases the expression of β-catenin and β-catenin-mediated cyclin D1, c-Myc, and PKM2 expression, thereby leading to enhanced aerobic glycolysis, ESCC cell proliferation, and tumour formation in mice. In addition, downregulated APC expression correlates with upregulated METTL3 expression in human ESCC specimens and poor prognosis in ESCC patients. Our findings reveal a mechanism by which the Wnt/β-catenin pathway is upregulated in ESCC via METTL3/YTHDF-coupled epitranscriptomal downregulation of APC.

The epitranscriptomic regulation of adenomatous polyposis coli (APC) tumour suppressor gene in cancers is unclear. Here the authors show that N6-methyladenosine methylation writer METTL3 downregulates APC by recruiting YTHDF2 for APC mRNA degradation, and that this promotes glycolysis and tumour growth in oesophageal cancers.

Conserved DNA sequence features underlie pervasive RNA polymerase pausing

Pausing of transcribing RNA polymerase is regulated and creates opportunities to control gene expression. Research in metazoans has so far mainly focused on RNA polymerase II (Pol II) promoter-proximal pausing leaving the pervasive nature of pausing and its regulatory potential in mammalian cells unclear. Here, we developed a pause detecting algorithm (PDA) for nucleotide-resolution occupancy data and a new native elongating transcript sequencing approach, termed nested NET-seq, that strongly reduces artifactual peaks commonly misinterpreted as pausing sites. Leveraging PDA and nested NET-seq reveal widespread genome-wide Pol II pausing at single-nucleotide resolution in human cells. Notably, the majority of Pol II pauses occur outside of promoter-proximal gene regions primarily along the gene-body of transcribed genes. Sequence analysis combined with machine learning modeling reveals DNA sequence properties underlying widespread transcriptional pausing including a new pause motif. Interestingly, key sequence determinants of RNA polymerase pausing are conserved between human cells and bacteria. These studies indicate pervasive sequence-induced transcriptional pausing in human cells and the knowledge of exact pause locations implies potential functional roles in gene expression.

The Emerging Roles of Cytosine-5 Methylation in mRNAs

Recent studies have unequivocally confirmed the presence of 5-methylcytosine (m5C) in mammalian mRNAs while indicating significant functional roles for this internal base modification type. Here, a brief history of m5C epitranscriptome research and a discussion of the important ways in which the field may now progress is presented.

Analysis of RNA Modifications by Second- and Third-Generation Deep Sequencing: 2020 Update

The precise mapping and quantification of the numerous RNA modifications that are present in tRNAs, rRNAs, ncRNAs/miRNAs, and mRNAs remain a major challenge and a top priority of the epitranscriptomics field. After the keystone discoveries of massive m⁶A methylation in mRNAs, dozens of deep sequencing-based methods and protocols were proposed for the analysis of various RNA modifications, allowing us to considerably extend the list of detectable modified residues. Many of the currently used methods rely on the particular reverse transcription signatures left by RNA modifications in cDNA; these signatures may be naturally present or induced by an appropriate enzymatic or chemical treatment. The newest approaches also include labeling at RNA abasic sites that result from the selective removal of RNA modification or the enhanced cleavage of the RNA ribose-phosphate chain (perhaps also protection from cleavage), followed by specific adapter ligation. Classical affinity/immunoprecipitation-based protocols use either antibodies against modified RNA bases or proteins/enzymes, recognizing RNA modifications. In this survey, we review the most recent achievements in this highly dynamic field, including promising attempts to map RNA modifications by the direct single-molecule sequencing of RNA by nanopores.

Methyladenosine Modification in RNAs: Classification and Roles in Gastrointestinal Cancers

Cellular ribonucleic acids (RNAs), including messenger RNAs (mRNAs) and non-coding RNAs (ncRNAs), harbor more than 150 forms of chemical modifications, among which methylation modifications are dynamically regulated and play significant roles in RNA metabolism. Recently, dysregulation of RNA methylation modifications is found to be linked to various physiological bioprocesses and many human diseases. Gastric cancer (GC) and colorectal cancer (CRC) are two main gastrointestinal-related cancers (GIC) and the most leading causes of cancer-related death worldwide. In-depth understanding of molecular mechanisms on GIC can provide important insights in developing novel treatment strategies for GICs. In this review, we focus on the multitude of epigenetic changes of RNA methlyadenosine modifications in gene expression, and their roles in GIC tumorigenesis, progression, and drug resistance, and aim to provide the potential therapeutic regimens for GICs.

Detection of genetic variation and base modifications at base-pair resolution on both DNA and RNA

Accurate decoding of nucleic acid variation is critical to understand the complexity and regulation of genome function. Here we use a single-molecule magnetic tweezer (MT) platform to identify sequence variation and map a range of important epigenetic base modifications with high sensitivity, specificity, and precision in the same single molecules of DNA or RNA. We have also developed a highly specific amplification-free CRISPR-Cas enrichment strategy to isolate genomic regions from native DNA. We demonstrate enrichment of DNA from both E. coli and the FMR1 5'UTR coming from cells derived from a Fragile X carrier. From these kilobase-length enriched molecules we could characterize the differential levels of adenine and cytosine base modifications on E. coli, and the repeat expansion length and methylation status of FMR1. Together these results demonstrate that our platform can detect a variety of genetic, epigenetic, and base modification changes concomitantly within the same single molecules.

Deciphering RNA modifications at base resolution: from chemistry to biology

Nearly 200 distinct chemical modifications of RNAs have been discovered to date. Their analysis via direct methods has been possible in abundant RNA species, such as ribosomal, transfer or viral RNA, since several decades. However, their analysis in less abundant RNAs species, especially cellular messenger RNAs, was rendered possible only recently with the advent of high throughput sequencing techniques. Given the growing biomedical interest of the proteins that write, erase and read RNA modifications, ingenious new methods to enrich and identify RNA modifications at base resolution have been implemented, and more efforts are underway to render them more quantitative. Here, we review several crucial modification-specific (bio)chemical approaches and discuss their advantages and shortcomings for exploring the epitranscriptome.

The nature of the modification at position 37 of tRNAPhe correlates with acquired taxol resistance

Acquired drug resistance is a major obstacle in cancer therapy. Recent studies revealed that reprogramming of tRNA modifications modulates cancer survival in response to chemotherapy. However, dynamic changes in tRNA modification were not elucidated. In this study, comparative analysis of the human cancer cell lines and their taxol resistant strains based on tRNA mapping was performed by using UHPLC-MS/MS. It was observed for the first time in all three cell lines that 4-demethylwyosine (imG-14) substitutes for hydroxywybutosine (OHyW) due to tRNA-wybutosine synthesizing enzyme-2 (TYW2) downregulation and becomes the predominant modification at the 37th position of tRNAphe in the taxol-resistant strains. Further analysis indicated that the increase in imG-14 levels is caused by downregulation of TYW2. The time courses of the increase in imG-14 and downregulation of TYW2 are consistent with each other as well as consistent with the time course of the development of taxol-resistance. Knockdown of TYW2 in HeLa cells caused both an accumulation of imG-14 and reduction in taxol potency. Taken together, low expression of TYW2 enzyme promotes the cancer survival and resistance to taxol therapy, implying a novel mechanism for taxol resistance. Reduction of imG-14 deposition offers an underlying rationale to overcome taxol resistance in cancer chemotherapy.

It Takes NSUN2 to Beat the Heat in Rice

Cytosine methylation (m⁵C) exists at low but significant levels in eukaryotic mRNAs, but its biological functions remain elusive. In this issue of Developmental Cell, Tang et al. find that the RNA methyltransferase OsNSUN2 is required for heat tolerance in rice by enhancing translation of specific transcripts under heat stress.

Enhanced differentiation of isomeric RNA modifications by reducing the size of ions in ion mobility mass spectrometric measurements

With the ability to differentiate different molecular sizes, ion mobility spectrometry (IMS) has great potentials in the analysis of isomeric compounds. However, due to the lack of sensitivity and resolution, IMS has not been commonly used. To address the issue on resolution, the goals of this study are to explore a more effective way to perform IMS by reducing the size of ions prior to the IM measurements, and apply the new approach to the differentiation of isomeric RNA modifications. The size reduction of ribonucleoside ions was effectively accomplished by using the collision-induced dissociation process, in which the N-glycosidic bond in ribonucleoside was cleaved and split the ions into two parts—a smaller nucleobase ion and a neutral molecule of ribose sugar. Since the chemical group that corresponds to most of the RNA modifications makes up a relatively small part of the molecular structure of nucleobases, the differentiation of the dissociated nucleobase ions is expected to require a lower ion mobility resolution than the differentiation of bigger isomeric ribonucleoside ions. By using RNA methylation as a model in this study, the proposed method lowered the required resolution by 16% for the differentiation of 1-methyladenosine and N6-methyladenosine. Similar results were also obtained from the differentiation of methylated cytidine isomers. In comparison to the results obtained from using the conventional tandem mass spectrometric method, there was no significant loss of signals when the proposed method was used. The proposed method is expected to be applicable to other types of isomeric compounds. Also, the same approach is applicable on other IMS platforms.

The epitranscriptome beyond m6A

Following its transcription, RNA can be modified by >170 chemically distinct types of modifications - the epitranscriptome. In recent years, there have been substantial efforts to uncover and characterize the modifications present on mRNA, motivated by the potential of such modifications to regulate mRNA fate and by discoveries and advances in our understanding of N ⁶-methyladenosine (m⁶A). Here, we review our knowledge regarding the detection, distribution, abundance, biogenesis, functions and possible mechanisms of action of six of these modifications - pseudouridine (Ψ), 5-methylcytidine (m⁵C), N ¹-methyladenosine (m¹A), N ⁴-acetylcytidine (ac⁴C), ribose methylations (N_m) and N ⁷-methylguanosine (m⁷G). We discuss the technical and analytical aspects that have led to inconsistent conclusions and controversies regarding the abundance and distribution of some of these modifications. We further highlight shared commonalities and important ways in which these modifications differ with respect to m⁶A, based on which we speculate on their origin and their ability to acquire functions over evolutionary timescales.

HydraPsiSeq: a method for systematic and quantitative mapping of pseudouridines in RNA

Developing methods for accurate detection of RNA modifications remains a major challenge in epitranscriptomics. Next-generation sequencing-based mapping approaches have recently emerged but, often, they are not quantitative and lack specificity. Pseudouridine (ψ), produced by uridine isomerization, is one of the most abundant RNA modification. ψ mapping classically involves derivatization with soluble carbodiimide (CMCT), which is prone to variation making this approach only semi-quantitative. Here, we developed 'HydraPsiSeq', a novel quantitative ψ mapping technique relying on specific protection from hydrazine/aniline cleavage. HydraPsiSeq is quantitative because the obtained signal directly reflects pseudouridine level. Furthermore, normalization to natural unmodified RNA and/or to synthetic in vitro transcripts allows absolute measurements of modification levels. HydraPsiSeq requires minute amounts of RNA (as low as 10-50 ng), making it compatible with high-throughput profiling of diverse biological and clinical samples. Exploring the potential of HydraPsiSeq, we profiled human rRNAs, revealing strong variations in pseudouridylation levels at ∼20-25 positions out of total 104 sites. We also observed the dynamics of rRNA pseudouridylation throughout chondrogenic differentiation of human bone marrow stem cells. In conclusion, HydraPsiSeq is a robust approach for the systematic mapping and accurate quantification of pseudouridines in RNAs with applications in disease, aging, development, differentiation and/or stress response.

METTL4 catalyzes m6Am methylation in U2 snRNA to regulate pre-mRNA splicing

N⁶-methylation of 2′-O-methyladenosine (Am) in RNA occurs in eukaryotic cells to generate N⁶,2′-O-dimethyladenosine (m⁶Am). Identification of the methyltransferase responsible for m⁶Am catalysis has accelerated studies on the function of m⁶Am in RNA processing. While m⁶Am is generally found in the first transcribed nucleotide of mRNAs, the modification is also found internally within U2 snRNA. However, the writer required for catalyzing internal m⁶Am formation had remained elusive. By sequencing transcriptome-wide RNA methylation at single-base-resolution, we identified human METTL4 as the writer that directly methylates Am at U2 snRNA position 30 into m⁶Am. We found that METTL4 localizes to the nucleus and its conserved methyltransferase catalytic site is required for U2 snRNA methylation. By sequencing human cells with overexpressed Mettl4, we determined METTL4’s in vivo target RNA motif specificity. In the absence of Mettl4 in human cells, U2 snRNA lacks m⁶Am thereby affecting a subset of splicing events that exhibit specific features such as 3′ splice-site weakness and an increase in exon inclusion. These findings suggest that METTL4 methylation of U2 snRNA regulates splicing of specific pre-mRNA transcripts.

A molecular-level perspective on the frequency, distribution, and consequences of messenger RNA modifications

Cells use chemical modifications to alter the sterics, charge, and conformations of large biomolecules, modulating their biogenesis, function, and stability. Until recently post-transcriptional RNA modifications were thought to be largely limited to nonprotein coding RNA species. However, this dogma has rapidly transformed with the discovery of a host of modifications in protein coding messenger RNAs (mRNAs). Recent advancements in genome-wide sequencing technologies have enabled the identification of mRNA modifications as a potential new frontier in gene regulation-leading to the development of the epitranscriptome field. As a result, there has been a flurry of multiple groundbreaking discoveries, including new modifications, nucleoside modifying enzymes ("writers" and "erasers"), and RNA binding proteins that recognize chemical modifications ("readers"). These discoveries opened the door to understanding how post-transcriptional mRNA modifications can modulate the mRNA lifecycle, and established a link between the epitranscriptome and human health and disease. Despite a rapidly growing recognition of their importance, fundamental questions regarding the identity, prevalence, and functional consequences of mRNA modifications remain to be answered. Here, we highlight quantitative studies that characterize mRNA modification abundance, frequency, and interactions with cellular machinery. As the field progresses, we see a need for the further integration of quantitative and reductionist approaches to complement transcriptome wide studies in order to establish a molecular-level framework for understanding the consequences of mRNA chemical modifications on biological processes. This article is categorized under: RNA Structure and Dynamics > RNA Structure, Dynamics and Chemistry RNA Processing > RNA Editing and Modification.

Detection of internal N7-methylguanosine (m7G) RNA modifications by mutational profiling sequencing

Methylation of guanosine on position N7 (m⁷G) on internal RNA positions has been found in all domains of life and have been implicated in human disease. Here, we present m⁷G Mutational Profiling sequencing (m⁷G-MaP-seq), which allows high throughput detection of m⁷G modifications at nucleotide resolution. In our method, m⁷G modified positions are converted to abasic sites by reduction with sodium borohydride, directly recorded as cDNA mutations through reverse transcription and sequenced. We detect positions with increased mutation rates in the reduced and control samples taking the possibility of sequencing/alignment error into account and use replicates to calculate statistical significance based on log likelihood ratio tests. We show that m⁷G-MaP-seq efficiently detects known m⁷G modifications in rRNA with mutational rates up to 25% and we map a previously uncharacterised evolutionarily conserved rRNA modification at position 1581 in Arabidopsis thaliana SSU rRNA. Furthermore, we identify m⁷G modifications in budding yeast, human and arabidopsis tRNAs and demonstrate that m⁷G modification occurs before tRNA splicing. We do not find any evidence for internal m⁷G modifications being present in other small RNA, such as miRNA, snoRNA and sRNA, including human Let-7e. Likewise, high sequencing depth m⁷G-MaP-seq analysis of mRNA from E. coli or yeast cells did not identify any internal m⁷G modifications.

Naturally occurring modified ribonucleosides

The chemical identity of RNA molecules beyond the four standard ribonucleosides has fascinated scientists since pseudouridine was characterized as the “fifth” ribonucleotide in 1951. Since then, the ever‐increasing number and complexity of modified ribonucleosides have been found in viruses and throughout all three domains of life. Such modifications can be as simple as methylations, hydroxylations, or thiolations, complex as ring closures, glycosylations, acylations, or aminoacylations, or unusual as the incorporation of selenium. While initially found in transfer and ribosomal RNAs, modifications also exist in messenger RNAs and noncoding RNAs. Modifications have profound cellular outcomes at various levels, such as altering RNA structure or being essential for cell survival or organism viability. The aberrant presence or absence of RNA modifications can lead to human disease, ranging from cancer to various metabolic and developmental illnesses such as Hoyeraal–Hreidarsson syndrome, Bowen–Conradi syndrome, or Williams–Beuren syndrome. In this review article, we summarize the characterization of all 143 currently known modified ribonucleosides by describing their taxonomic distributions, the enzymes that generate the modifications, and any implications in cellular processes, RNA structure, and disease. We also highlight areas of active research, such as specific RNAs that contain a particular type of modification as well as methodologies used to identify novel RNA modifications.

This article is categorized under:

RNA Processing > RNA Editing and Modification

Direct RNA sequencing enables m6A detection in endogenous transcript isoforms at base-specific resolution

Direct RNA sequencing holds great promise for the de novo identification of RNA modifications at single-coordinate resolution; however, interpretation of raw sequencing output to discover modified bases remains a challenge. Using Oxford Nanopore's direct RNA sequencing technology, we developed a random forest classifier trained using experimentally detected N⁶-methyladenosine (m⁶A) sites within DRACH motifs. Our software MINES (m⁶A Identification using Nanopore Sequencing) assigned m⁶A methylation status to more than 13,000 previously unannotated DRACH sites in endogenous HEK293T transcripts and identified more than 40,000 sites with isoform-level resolution in a human mammary epithelial cell line. These sites displayed sensitivity to the m⁶A writer, METTL3, and eraser, ALKBH5, respectively. MINES (https://github.com/YeoLab/MINES.git) enables m⁶A annotation at single coordinate-level resolution from direct RNA nanopore sequencing.

[Long non-coding RNAs show different expression profiles and display competing endogenous RNA potential in placenta accreta spectrum disorders]

OBJECTIVE

To investigate the expression profile of long non-coding RNAs (lncRNA) and identify potential lncRNA-related competing endogenous RNAs (ceRNA) in placenta accrete spectrum disorders (PAS).

METHODS

Five tissue specimens of placental implantation and 5 adjacent normal placental tissues were collected from cesarean section deliveries complicated by PAS in our hospital between December, 2017 and June, 2018. Human microarrays were used to identify the lncRNAs that were differentially expressed in PAS, and 5 of the identified lncRNAs were further validated using qRT-PCR. GO and KEGG pathway analyses were performed to indentify the most significant enrichment functions. A ceRNA network was constructed based on ENST00000511361 (RP5-875H18.4), NR_027457 (LINC00221) and NR_126415 (FOXP4-AS1) to pinpoint the potential lncRNAs-related ceRNA.

RESULTS

A total of 329 lncRNAs and 179 mRNAs were identified to have differential expression in PAS. The results of qRT-PCR were consistent with the human microarrays results. Transforming growth factor-β (TGF-β) signaling pathway was the most significantly enriched pathway. The constructed ceRNA network suggested that RP5-875H18.4--miRNA-218--SLIT2 had a potential ceRNA regulatory mechanism in PAS.

CONCLUSIONS

The differentially expressed lncRNAs are involved in the occurrence and progression of PAS possibly by regulating the TGF-β signaling pathway. The ceRNA network of RP5-875H18.4--miRNA-218--SLIT2 may play a role in the occurrence of PAS.

HLMethy: a machine learning-based model to identify the hidden labels of m6A candidates

We developed a machine learning-based model to identify the hidden labels of m⁶A candidates from noisy m6A-seq data. Peak-calling approaches, such as MeRIP-seq or m⁶A-seq, are commonly used to map m⁶A modifications. However, these technologies can only map m⁶A sites with 100-200 nt resolution and cannot reveal the precise location or the number of modified residues in a transcript. To address this challenge, we developed a novel machine learning-based approach, named HLMethy, to assign labels to m⁶A candidates from noisy m⁶A-seq data. The multiple instance learning framework was adopted and two different training strategies were used to generate the classification model. To test the performance of our model, the m⁶A sites with single-base resolution were used and our model achieved comparable performance against existing instance-level predictors, which suggest that our model has the potential to improve the data quality of m⁶A-seq at reduced costs. What's more, our generic framework can be extended to other newly found modifications that are found by peak-calling approaches. The source code of HLMethy is available at https://github.com/liuze-nwafu/HLMethy.

Structural Elucidation of Bisulfite Adducts to Pseudouridine That Result in Deletion Signatures during Reverse Transcription of RNA

The recent report of RBS-Seq to map simultaneously the epitranscriptomic modifications N¹-methyladenosine, 5-methylcytosine, and pseudouridine (Ψ) via bisulfite treatment of RNA provides a key advance to locate these important modifications. The locations of Ψ were found by a deletion signature generated during cDNA synthesis after bisulfite treatment for which the chemical details of the reaction are poorly understood. In the present work, the bisulfite reaction with Ψ was explored to identify six isomers of bisulfite adducted to Ψ. We found four of these adducts involved the heterocyclic ring, similar to the reaction with other pyrimidines. The remaining two adducts were bonded to the 1' carbon, which resulted in opening of the ribose ring. The utilization of complementary 1D- and 2D-NMR, Raman, and electronic circular dichroism spectroscopies led to the assignment of the two ribose adducts being the constitutional isomers of an S- and an O-adduct of bisulfite to the ribose, and these are the final products after heating. A mechanistic proposal is provided to rationalize chemically the formation and stereochemistries of all six isomeric bisulfite adducts to Ψ; conversion of intermediate adducts to the two final products is proposed to involve E2, S_N2', and [2,3]-sigmatropic shift reactions. Lastly, a synthetic RNA template with Ψ at a known location was treated with bisulfite, leading to a deletion signature after reverse transcription, supporting the RBS-Seq report. This classical bisulfite reaction used for epigenomic and epitranscriptomic sequencing diverges from the C nucleoside Ψ to form stable bisulfite end products that yield signatures for next-generation sequencing.