Sarah Rennie

Direct RNA nanopore sequencing offers the capacity to directly profile native RNAs, enabling detection of diverse epitranscriptomic modifications based on current changes as the RNA passes through a pore. Innovations in chemistry, hardware and bioinformatic tools have increased sequencing capacity and accuracy, creating demand for sequencing multiple samples within individual runs. Two recent solutions enable barcoding and demultiplexing of samples, but independent comparisons of their performance is limited. Here we benchmark WarpDemuX and SeqTagger using orthogonal ground truths derived from species identity and matched IVT spike-ins across four RNA004 flow cells. Both achieve high demultiplexing precision but differ in their trade-offs in precision and retention. SeqTagger provided strong performance at its recommended settings, whilst WarpDemuX performance varied by model, barcode and read quality. IVT-derived score cutoffs allowed dataset-specific calibration and increased WarpDeMux assignment yield, although precision achieved on IVTs was not consistently reproduced in species reads. These results support matched IVTs as internal controls for guiding choice of demultiplexing settings, while demonstrating the need to validate demultiplexing precision in biological RNA populations.

Human papillomavirus (HPV)-driven cancers remain a major global health burden, and understanding how post-transcriptional regulation shapes viral gene expression may inform new therapeutic strategies. Here, we reanalysed independently generated public HeLa datasets to identify candidate RNA modification sites on integrated HPV18 transcripts expressed from the HeLa genome. Using Oxford Nanopore direct RNA sequencing of native RNAs and in vitro transcribed controls, together with GLORI, eTAM-seq and staged 4sU-GLORI datasets, we identified a set of candidate m6A sites on HPV18 early transcripts. m6A levels at a subset of these sites were reduced following perturbation of the m6A pathway, most clearly after METTL3 inhibition, WTAP knockdown or FTO overexpression. We further show that the E6*I-proximal m6A site at position 224 is enriched on unspliced transcripts in direct RNA sequencing, eTAM-seq and staged 4sU-GLORI data. In contrast, we find no convincing evidence for m5C or pseudouridine within the HPV18 regions covered by these datasets. Together, our analyses provide a prioritized candidate map of RNA modifications on HeLa-expressed HPV18 early transcripts.

PUMILIO/Fem-3 Binding Factor (PUF) proteins are conserved eukaryotic mRNA-binding proteins. Higher plants encode an expanded family of PUF proteins with 26 members in Arabidopsis thaliana of which 6 (PUM1-PUM6) have a domain organization equivalent to metazoan PUF proteins implicated in mRNA regulation. Here, we show that the PUF proteins PUM2 and PUM6 are expressed in mitotically active cells of root and shoot meristems, while PUM5 is mostly expressed in the differentiation zone of root meristems exiting from active division. Increased PUM2 dosage in proliferating cells causes growth repression and developmental delay reminiscent of mutants with defects in cytoplasmic YTHDF proteins that bind to N6-methyladenosine (m6A) in mRNA. Remarkably, even weak defects in the m6A-YTHDF system enhance the effect of increased PUM2 dosage. In line with this genetic interaction, mRNA targets overlap significantly between PUFs and the major YTHDF proteins ECT2 and ECT3. Furthermore, the enrichment of the core PUF recognition motif URUAY around m6A sites is largely explained by PUF targeting, not by direct URUAY methylation as previously suggested. Our results indicate that the balance between the reciprocal activities of the deeply conserved cytoplasmic RNA binding protein families PUF and YTHDF is an important element of growth control in plants.

Human papillomavirus (HPV)-driven cancers remain a major global health burden, and understanding how post-transcriptional regulation shapes viral gene expression may inform new therapeutic strategies. Here, we reanalysed independently generated public HeLa datasets to identify candidate RNA modification sites on integrated HPV18 transcripts expressed from the HeLa genome. Using Oxford Nanopore direct RNA sequencing of native RNAs and in vitro transcribed controls, together with GLORI, eTAM-seq and staged 4sU-GLORI datasets, we identified a set of candidate m ⁶ A sites on HPV18 early transcripts. m ⁶ A levels at a subset of these sites were reduced following perturbation of the m ⁶ A pathway, most clearly after METTL3 inhibition, WTAP knockdown or FTO overexpression. We further show that the E6*I-proximal m ⁶ A site at position 224 is enriched on unspliced transcripts in direct RNA sequencing, eTAM-seq and staged 4sU-GLORI data. In contrast, we find no convincing evidence for m5C or pseudouridine within the HPV18 regions covered by these datasets. Together, our analyses provide a candidate map of RNA modifications on HeLa-expressed HPV18 early transcripts.

Transcription factor (TF) cooperativity plays a critical role in gene regulation. However, the underlying genomic rules remain unclear, calling for scalable methods to characterize the TF binding site (motif) syntax of regulatory elements. Here, we introduce DeepCompARE, a lightweight model paired with an in silico ablation (ISA) framework for genome-wide analysis of regulatory sequences. Our framework enables precise interpretation of the motif syntax governing chromatin accessibility, enhancer activity, and promoter function. We find that most TF motifs are pairwise independent, indicating a default additive behavior of TFs, and define a cooperativity score to quantify deviations from this baseline. This reveals synergy and redundancy as opposite effects along the same cooperative spectrum. TF redundancy is linked to promoter activity and broad expression, whereas TF synergy is associated with enhancer activity, physical interactions, and cell-type specificity. Our framework provides a quantitative model for TF cooperativity, offering new insights into gene regulatory logic.

N6-methyladenosine (m ⁶ A) exerts many of its regulatory effects on eukaryotic mRNAs by recruiting cytoplasmic YT521-B homology-domain family (YTHDF) proteins. Here, we show that in Arabidopsis thaliana , the interaction between m ⁶ A and the major YTHDF protein ECT2 also involves the mRNA-binding ALBA protein family. ALBA and YTHDF proteins physically associate via a deeply conserved short linear motif in the intrinsically disordered region of YTHDF proteins and their mRNA target sets overlap, with ALBA4 binding sites being juxtaposed to m ⁶ A sites. These binding sites correspond to pyrimidine-rich elements previously found to be important for m ⁶ A binding to ECT2. Accordingly, both the biological functions of ECT2, and its binding to m ⁶ A targets in vivo, require ALBA association. Our results introduce the YTHDF-ALBA complex as the functional cytoplasmic m ⁶ A-reader in Arabidopsis , and define a molecular foundation for the concept of facilitated m ⁶ A reading, which increases the potential for combinatorial control of biological m ⁶ A effects.

RNA-binding proteins and chemical modifications to RNA play vital roles in the co- and post-transcriptional regulation of genes. In order to fully decipher their biological roles, it is an essential task to catalogue their precise target locations along with their preferred contexts and sequence-based determinants. Recently, deep learning approaches have significantly advanced in this field. These methods can predict the presence or absence of modification at specific genomic regions based on diverse features, particularly sequence and secondary structure, allowing us to decipher the highly non-linear sequence patterns and structures that underlie site preferences. This article provides an overview of how deep learning is being applied to this area, with a particular focus on the problem of mRNA-RBP binding, while also considering other types of chemical modification to RNA. It discusses how different types of model can handle sequence-based and/or secondary-structure-based inputs, the process of model training, including choice of negative regions and separating sets for testing and training, and offers recommendations for developing biologically relevant models. Finally, it highlights four key areas that are crucial for advancing the field.

Long non-coding RNAs (lncRNAs) typically refer to non-protein coding RNAs that are longer than 200 nucleotides. Historically dismissed as junk DNA, over two decades of research have revealed that lncRNAs bind to other macromolecules (e.g., DNA, RNA, and/or proteins) to modulate signaling pathways and maintain organism viability. Their discovery has been significantly aided by the development of bioinformatics tools in recent years. However, the diversity of tools for lncRNA discovery and functional prediction can present a challenge for researchers, especially bench scientists and clinicians. This Perspective article aims to navigate the current landscape of bioinformatic tools suitable for both protein-coding and lncRNA genes. It aims to provide a guide for bench scientists and clinicians to select the appropriate tools for their research questions and experimental designs.

RNA, like DNA and proteins, can undergo modifications. To date, over 170 RNA modifications have been identified, leading to the emergence of a new research area known as epitranscriptomics. RNA editing is the most frequent RNA modification in mammalian transcriptomes, and two types have been identified: (1) the most frequent, adenosine to inosine (A-to-I); and (2) the less frequent, cysteine to uracil (C-to-U) RNA editing. Unlike other epitranscriptomic marks, RNA editing can be readily detected from RNA sequencing (RNA-seq) data without any chemical conversions of RNA before sequencing library preparation. Furthermore, analyzing RNA editing patterns from transcriptomic data provides an additional layer of information about the epitranscriptome. As the significance of epitranscriptomics, particularly RNA editing, gains recognition in various fields of biology and medicine, there is a growing interest in detecting RNA editing sites (RES) by analyzing RNA-seq data. To cope with this increased interest, several bioinformatic tools are available. However, each tool has its advantages and disadvantages, which makes the choice of the most appropriate tool for bench scientists and clinicians difficult. Here, we have benchmarked bioinformatic tools to detect RES from RNA-seq data. We provide a comprehensive view of each tool and its performance using previously published RNA-seq data to suggest recommendations on the most appropriate for utilization in future studies.

The NLRP3 inflammasome plays a pivotal role in regulating inflammation and immune responses. Its activation can lead to an inflammatory response and pyroptotic cell death. This is beneficial in the case of infections, but excessive activation can lead to chronic inflammation and tissue damage. Moreover, while most of the mammalian genome is transcribed as RNAs, only a small fraction codes for proteins. Among non-protein-coding RNAs, long non-coding RNAs (lncRNAs) have been shown to play key roles in regulating gene expression and cellular processes. They interact with DNA, RNAs, and proteins, and their dysregulation can provide insights into disease mechanisms, including NLRP3 inflammasome activation. Here, we systematically analyzed previously published RNA sequencing (RNA-seq) data of NLRP3 inflammasome activation in monocytes/macrophages to uncover inflammasome-regulated lncRNA genes. To uncover the functional importance of inflammasome-regulated lncRNA genes, one inflammasome-regulated lncRNA, ENSG00000273124, was knocked down in an in vitro model of macrophage polarization. The results indicate that silencing of ENSG00000273124 resulted in the up-regulation tumor necrosis factor (TNF), suggesting that this lncRNA might be involved in pro-inflammatory response in macrophages. To make our analyzed data more accessible, we developed the web database InflammasomeDB.