Biology subjects

Capitanchik, C.

Publications and source records attributed to Capitanchik, C..

8 recordsLinked to original sources

bioRxiv2026

PARNET: A CLIP-SEQ-BASED FOUNDATION MODEL FOR RNA SEQUENCE REPRESENTATION LEARNING

RNA-binding proteins (RBPs) orchestrate a complex combinatorial regulatory "code" that governs RNA splicing, stability, localization, and translation. Learning the relationship between RNA sequences and these processes is a central challenge in genomics. Foundation models, notably RNA language models, have emerged as the dominant approach, learning general-purpose representations from unlabeled sequence at scale. While RNA language models have demonstrated impressive performance across a broad range of downstream tasks, they generally learn from sequence reconstruction objectives alone, lacking direct connections to the regulatory principles that govern RNA function. Here we introduce Parnet, an RNA foundation model trained directly and exclusively on experimental CLIP-seq data. Parnet is a multi-task foundation model trained end-to-end on 223 eCLIP-seq experiments spanning 150 RBPs to predict base-resolution RBP binding profiles directly from RNA sequence. This CLIP-seq pretraining strategy departs fundamentally from the masked-language-modeling paradigm, anchoring learned RNA representations directly in measured protein-RNA interactions rather than sequence statistics. Parnet substantially outperforms its single-task predecessor RBPNet in binding profile and motif recovery, generalizes to unseen cell types and iCLIP data, and recapitulates position-dependent splicing regulation. Frozen Parnet embeddings, without task-specific fine-tuning, match or exceed the performance of both task-specific tools, as well as larger self-supervised RNA and genomic language models across diverse downstream tasks, including RNA biotype classification, lncRNA chromatin localization, translational efficiency, splice-site recognition, intron retention, and non-coding variant effect prediction. Importantly, Parnet remains mechanistically interpretable, tracing predictions back to the specific RBPs and motifs that drive them. These results establish the RBP interactome as a compact, functionally sufficient, and interpretable basis for foundation model pretraining in RNA biology.

bioinformatics↗

bioRxiv2025

PRMT activity promotes global 3' UTR shortening in proliferating cells

Protein methyltransferase (PRMT)-catalysed arginine methylation is a widespread post-translational modification that regulates numerous RNA-binding proteins and frequently becomes dysregulated in cancer. While PRMT inhibitors have shown promise as an anti-cancer strategy, greater understanding of the downstream pathways linking arginine methylation to tumour-promoting phenotypes is needed to improve patient stratification and develop more effective therapeutic approaches. Here, we reveal arginine methylation as a critical regulator of alternative polyadenylation (APA) patterns that are fundamental to tumour progression. 3' RNA-sequencing assays uncover a rapid and global shift toward longer 3' UTR isoforms upon dual (symmetric and asymmetric) methylation (DMAi), impacting a broad range of cellular proliferation and signalling genes. Arginine methylation is required for sustaining proximal poly(A) site usage under high proliferative demand, as DMAi treatment blocks use of such sites in activated T cells, various cancer cell lines and patient-derived lung organoids. DMAi also counteracts the 3' UTR shortening caused by reduced CFIM25 expression, which normally promotes oncogenic isoforms. DMAi treatment affects APA in many of the same mRNAs as impaired cleavage and polyadenylation activity, and these mRNAs contain characteristic signatures such as high GC-content and long 3 UTRs. This systematic impact of PRMT activity on APA regulation broadens the potential utility of PRMT inhibitors as therapeutic agents for both cancer and immune-related diseases.

molecular biology↗

bioRxiv2023

Flow: a web platform and open database to analyse, store, curate and share bioinformatics data at scale

Ever-increasing volumes of sequencing data offer potential for large-scale meta-analyses to address significant biological questions. However, challenges such as insufficient data processing information, data quality concerns, and issues related to accessibility and curation often present obstacles. Additionally, most experimental biologists lack the time and expertise needed to independently analyse, manage and share their own data. To overcome these hurdles, we present Flow, a web-based platform that links bioinformatic analysis and database solutions with a user-friendly interface and web API. Flow currently accommodates a range of genomics methods and further DSL2-compliant Nextflow pipelines can be added via a simple JSON schema file. Deployable on local systems or cloud services, an instance is freely accessible to academic researchers at https://flow.bio.

bioinformatics↗

bioRxiv2022

m6A-ELISA, a simple method for quantifying N6-methyladenosine from mRNA populations

N6-methyladenosine (m6A) is a widely studied and abundant RNA modification. The m6A mark regulates the fate of RNAs in various ways, which in turn, drives changes in cell physiology, development, and disease pathology. Over the last decade, numerous methods have been developed to map and quantify m6A sites genomewide through deep sequencing. Alternatively, m6A levels can be quantified from a population of RNAs using techniques such as liquid chromatography-mass spectrometry or thin layer chromatography. However, many methods for quantifying m6A levels involve extensive protocols and specialized data analysis, and often only a few samples can be handled in a single experiment. Here, we developed a simple method for determining m6A levels in mRNA populations from various sources based on enzyme-linked immunosorbent-based assay (m6A-ELISA). We have optimized various steps of m6A-ELISA such as sample preparation and the background signal resulting from the primary antibody. We validated the method using mRNA populations from budding yeast and mouse embryonic stem cells. The full protocol takes less than a day, requiring only 25 ng of mRNA. The m6A-ELISA protocol is therefore quick, cost-effective, and scalable, making it a valuable tool for determining relative m6A levels in samples from various sources that could be adapted to detect other mRNA modifications.

biochemistry↗

bioRxiv2022

m6A reader Pho92 is recruited co-transcriptionally and couples translation efficacy to mRNA decay to promote meiotic fitness in yeast

N6-methyladenosine (m6A) RNA modification impacts mRNA fate primarily via reader proteins, which dictate processes in development, stress, and disease. Yet little is known about m6A function in Saccharomyces cerevisiae, which occurs solely during early meiosis. Here we perform a multifaceted analysis of the m6A reader protein Pho92/Mrb1. Cross-linking immunoprecipitation analysis reveals that Pho92 associates with the 3end of meiotic mRNAs in both an m6A-dependent and independent manner. Within cells, Pho92 transitions from the nucleus to the cytoplasm, and associates with translating ribosomes. In the nucleus Pho92 associates with target loci through its interaction with transcriptional elongator Paf1C. Functionally, we show that Pho92 promotes and links protein synthesis to mRNA decay. As such, the Pho92-mediated m6A-mRNA decay is contingent on active translation and the CCR4-NOT complex. We propose that the m6A reader Pho92 is loaded co-transcriptionally to facilitate protein synthesis and subsequent decay of m6A modified transcripts, and thereby promotes meiosis.

biochemistry↗

bioRxiv2021

clipplotr - a comparative visualisation and analysis tool for CLIP data

CLIP technologies are now widely used to study RNA-protein interactions and many datasets are now publicly available. An important first step in CLIP data exploration is the visual inspection and assessment of processed genomic data on selected genes or regions and performing comparisons: either across conditions within a particular project, or incorporating publicly available data. However, the output files produced by data processing pipelines or preprocessed files available to download from data repositories are often not suitable for direct comparison and usually need further processing. Furthermore, to derive biological insight it is usually necessary to visualise CLIP signal alongside other data such as annotations, or orthogonal functional genomic data (e.g. RNA-seq). We have developed a simple, but powerful, command-line tool: clipplotr, which facilitates these visual comparative and integrative analyses with normalisation and smoothing options for CLIP data and the ability to show these alongside reference annotation tracks and functional genomic data. These data can be supplied as input to clipplotr in a range of file formats, which will output a publication quality figure. It is written in R and can both run on a laptop computer independently, or be integrated into computational workflows on a high-performance cluster. Releases, source code and documentation are freely available at: https://github.com/ulelab/clipplotr.

bioinformatics↗

bioRxiv2020

psiCLIP reveals dynamic RNA binding by DEAH-box helicases before and after exon ligation

Eight RNA helicases remodel the spliceosome to effect pre-mRNA splicing but their mechanism of action remains poorly understood. We have developed "purified spliceosome iCLIP" (psiCLIP) to define helicase-RNA contacts in specific spliceosomal states. psiCLIP reveals previously unappreciated dynamics of spliceosomal helicases. The binding profile of the helicase Prp16 is influenced by the distance between the branch-point and 3 splice site, while Prp22 binds diffusely on the intron before exon ligation but switches to more narrow binding downstream of the exon junction after exon ligation. Notably, depletion of the exon-ligation factor Prp18 destabilizes Prp22 binding to the pre-mRNA, demonstrating that psiCLIP can be used to study the relationships between helicases and auxiliary splicing factors. Thus, psiCLIP is sensitive to spliceosome dynamics and complements the insights from structural and imaging studies by providing crucial positional information on helicase-RNA contacts during spliceosomal remodeling.

molecular biology↗

bioRxiv2019

RNA modifications detection by comparative Nanopore direct RNA sequencing

RNA molecules undergo a vast array of chemical post-transcriptional modifications (PTMs) that can affect their structure and interaction properties. To date, over 150 naturally occurring PTMs have been identified, however the overwhelming majority of their functions remain elusive. In recent years, a small number of PTMs have been successfully mapped to the transcriptome using experimental approaches relying on high-throughput sequencing. Oxford Nanopore direct-RNA sequencing (DRS) technology has been shown to be sensitive to RNA modifications. We developed and validated Nanocompore, a robust analytical framework to evaluate the presence of modifications in DRS data. To do so, we compare an RNA sample of interest against a non-modified control sample. Our strategy does not require a training set and allows the use of replicates to model biological variability. Here, we demonstrate the ability of Nanocompore to detect RNA modifications at single-molecule resolution in human polyA+ RNAs, as well as in targeted non-coding RNAs. Our results correlate well with orthogonal methods, confirm previous observations on the distribution of N6-methyladenosine sites and provide novel insights into the distribution of RNA modifications in the coding and non-coding transcriptomes. The latest version of Nanocompore can be obtained at https://github.com/tleonardi/nanocompore.

genomics↗