RNAcentral Release 28

We are pleased to announce RNAcentral Release 28, which now contains more than 58 million sequences, an increase of approximately 3 million sequences since the previous release.

In this release we have focused on improving the release pipeline for both speed and correctness, with an aim to increase the number of releases happening within the year matching other major database cycles.

Notably, piRBase has been updated to v3.0 which includes curated gold standard sets. This import more than doubles the number of piRNA sequences and increases species coverage from 17 to 26. NONCODE and MGI were updated for the first time in several years leading to the addition of nearly 400,000 new lncRNA sequences and doubling the number of mouse sequences annotated by MGI. SILVA was also updated which adds over 2 million new small-subunit rRNA sequences.

Database updates

These databases were updated in this release to the version stated below:

ENA: 2026-05-15 -> 2026-09-17

WormBase 2026-05-15 -> 2026-09-17

Ensembl: 115 -> 116

Ensembl Fungi: 62 -> 63

Ensembl Metazoa: 62 -> 63

Ensembl Plants: 62 -> 63

Ensembl Protists: 62 -> 63

Ensembl/GENCODE: human 49/mouse M38 -> human 50/mouse M39

Expression Atlas: 2026-06-23 -> 2026-09-18

piRBase: v2 -> v3

PSIQUIC: 2021-09-03 -> 2026-09-17

HGNC: 2026-04-28 -> 2026-09-10

CRW: 2020-09-21 -> 2026-09-10

Ribovision: 2022-02-09 -> 2026-09-10

FlyBase: FB2026_01 -> FB2026_03

PomBase: 2026-04-28 -> 2026-09-09

RGD: 2026-04-28 -> 2026-09-09

ZFIN: 2026-03-27 -> 2026-09-09

RefSeq: v235 -> v237

SILVA: 138.2 -> 144

GeneCards: 5.26 -> 6.1

MalaCards: 5.26 -> 6.1

PDBe: 2026-04-29 -> 2026-09-09

QuickGo: 2026-04-13 -> 2026-09-09
NONCODE: v5 -> v7

MGI: 6.10 -> 6.24

R2DT updates

We have re-run all sequences so that they are drawn with version 2.2 or higher. We have also worked to correct diagrams in which a sequence had been drawn using the template for a different type of RNA.

Genome mapping updates

We have improved genome mapping to cover very large genomes, mostly plants and corrected our thresholds to map more sequences overall.

Website updates

DocBot

The RNAcentral website now has DocBot, an AI assistant that answers questions about RNAcentral using our documentation. You can ask it how to search the database, download data or use the API, and it will answer with links to the relevant help pages. It appears on the bottom right of the website.

GOFlowLLM annotations 

Some sequence pages show Gene Ontology annotations produced by GOFlowLLM. GOFlowLLM uses a large language model to follow a curation workflow over the scientific literature and suggest GO terms with supporting evidence. Where an RNA has been annotated by the system in multiple papers, these are now all shown on the RNA page, under the GoFlow Results tab. GoFlowLLM annotations are labelled so you can tell them apart from manually curated annotations. 

Public database schema changes

This release changes the schema of the public Postgres database. If you query it directly, please update your queries:

• Columns should now more accurately reflect their contents, e.g. the id column in rnc_rna_precomputed has been renamed urs_taxid

• The upi column is now called urs in the rna and xref tables, to match the URS identifiers used everywhere else in RNAcentral. A query like SELECT upi FROM rna becomes SELECT urs FROM rna.

• Some names that look old keep upi on purpose, such as a few index names. These don't affect queries. The FTP exports, API and website are not affected by these changes.

Get in touch

As always, all data are freely available on the RNAcentral website, via the API, our public database, and in the FTP archive. If you have any feedback, please get in touch by email, Bluesky, LinkedIn, or by submitting an issue on GitHub. We look forward to hearing from you!

RNAcentral release 27

We are pleased to announce RNAcentral Release 27, which now contains more than 55 million sequences, an increase of approximately 10 million sequences since the previous release.

For the first time, RNAcentral includes expert databases for circular RNAs (circRNAs), CIRCpedia and circAtlas. We also welcome mirtronDB, which provides mirtrons, a subclass of miRNAs derived from small introns, and JaponicusDB, which contributes RNA sequences from the fission yeast Schizosaccharomyces japonicus.

This release also expands RNAcentral's quality-control capabilities with two new tools for identifying potential protein-coding regions, tcode and stopfree, alongside a new QC summary table. These complement the existing CPAT, contamination, and Rfam quality-control checks.

We have expanded our RNA gene annotations to cover more than 800 species and introduced several improvements to the website, including a new track hub displaying Sci-ModoM RNA modifications in the IGV browser and data exports in Parquet format.

Data

New databases

JaponicusDB date imported 2026-06-23

CIRCpedia v3

circAtlas 3.0

Sci-ModoM (via MODOMICS) imported 2026-06

mirtronDB 2.0

Database updates

These databases were updated in this release to the version stated below:

ENA 2025-03-06 -> 2026-05-15

Rfam 15.0 -> 15.1

Ensembl 113 -> 115

Ensembl Fungi 60 -> 62

Ensembl Metazoa 60 -> 62

Ensembl Plants 60 -> 62

Ensembl Protists 60 -> 62

Ensembl/GENCODE human 47/mouse M34 -> human 49/mouse M38

Expression Atlas 2025-04-01 -> 2026-06-23

FlyBase FB2025_01 -> FB2026_01

GeneCards 5.23 -> 5.26

HGNC 2025-03-06 -> 2026-04-28

IntAct 2024-11-05 -> 2026-04-13

MalaCards 5.23 -> 5.26

PDBe 2024-11-05 -> 2026-04-29

PomBase 2024-11-05 -> 2026-04-28

QuickGo 2024-11-05 -> 2026-04-13 

RefSeq 2025-03-06 -> 2026-03-25 (v235)

SGD 2025-03-06 -> 2026-03-26

WormBase 2025-03-06 -> 2026-05-15

ZFIN 2024-11-05 -> 2026-03-27

TAIR 10 -> 12
gtRNAdb 21 -> 22
RGD March 2018 -> 2026-04-28
SILVA 138.1 -> 138.2
MGNIFY 2025-03-06 -> 2026-04-13
MODOMICS 2024 -> 2025
EVLNCRNAS 2.0 -> 3.0
REDIportal 2.0 -> 3.0

New gene-level entries

We have expanded the spectrum of species covered for genes, going from ~107k non-coding RNA genes (203 species, release 26) to a total of >1M genes (>800 species) in the current release.

AI-generated summaries for transcripts are now visible on the gene pages as well. For more information about gene-level entries, check our NAR 2026 paper.

Expanded coding-potential detection 

This release strengthens RNAcentral's quality control with two new coding-potential tools and a new summary table. Alongside the existing CPAT check, we now run tcode (EMBOSS v6.6.0), which applies the Fickett TESTCODE statistic across all organisms on sequences longer than 200 nt, and stopFree, an in-house Rust implementation that measures the longest stop-codon-free run in each reading frame, run on all sequences. To make these results easy to interpret, we have also introduced a quality-control summary table in the QC Status section, letting users assess a sequence at a glance. It reports coding potential, contamination, and Rfam checks, with a green tick for a pass and a red cross for a fail. The coding-potential row can be expanded to show the individual CPAT, tcode, and stopFree results, and displays an amber status when the three methods disagree. We acknowledge the inspiration from the benchmarking study Evaluating computational tools for protein-coding sequence detection: Are they up to the task? by Paul Gardner and colleagues, which informed our choice of coding-potential tools.

Litscan updates

Litscan has been updated to scan 23,032,875 ids across 1,509,944 papers. LitScan searches the scientific literature for mentions of RNAcentral identifiers and related terms (such as gene names, synonyms, and database accessions), surfacing publications relevant to each sequence.

We have extended LitScan's literature coverage to additional expert databases, including TarBase, LncBook, and DictyBase. By adding tailored queries for these databases, we improve the discovery of literature linked to their entries, for example matching gene synonyms such as LINC01786 for LncBook, or accessions and RNA names for TarBase and DictyBase.

RNAcentral MCP server

This release introduces the RNAcentral MCP Server, which exposes the RNAcentral database as tools that large language models can call directly. Built on the Model Context Protocol, it lets AI assistants query ncRNA sequences, map identifiers, find ncRNAs overlapping genomic regions, retrieve 2D structures, export sequences (FASTA or Parquet), and access literature summaries, orchestrating several upstream APIs so a single question is answered with one tool call.

It is published on PyPI as rnacentral-mcp-server, and listed in BioContextAI, a community effort to connect agentic AI with biomedical resources using the MCP. See the documentation for setup with Claude Desktop, Claude Code, and other clients.

Website updates

Sci-ModoM tracks with RNA modifications and highlighted transcripts in IGV

Sci-ModoM, the first next-generation RNome database offering a holistic view of the epitranscriptomic landscape. Sci-ModoM contains 7,952,319 reported sites with stoichiometric information across 163 datasets, for genome assemblies GRCh38 (human) and GRCm39 (mouse) [2026-06-15], provided in bedRMod format. We have made these RNA annotation variants available in a track hub in the IGV genome browser. In addition, finding the transcript you're viewing within the embedded IGV genome browser is now much easier. The transcript of interest is highlighted in yellow, so it stands out from the other transcripts in the same genomic region rather than blending in. This is a small change that makes navigating the genome browser on a transcript page noticeably quicker.

Example for human small nucleolar RNA host gene 29 (SNHG29)

https://rnacentral.org/rna/URS000075BC33/9606

Filtering in the Molecular Relationships tab

Browsing molecular relationships is now faster. We have added a Target filter to the Relationships tab, letting you narrow the table to the interactions you care about instead of paging through every entry. This is particularly helpful for RNAs with large numbers of relationships, where finding a specific target previously meant scrolling through long result sets.

Molecular relationships tab for human SNHG29. 

Export search results to Hugging Face in Parquet format

You can now export a set of search results directly as a Hugging Face dataset in Parquet format, ready for machine-learning workflows. From the export options, the chosen selection of sequences is packaged as Parquet (with an accompanying dataset README) and made available on the Hugging Face Hub, making it easy to build training or analysis datasets from RNAcentral without any manual conversion. Parquet is also offered as a standard download format for search exports.

Other web updates

• Faster sequence search: Sequence search now runs on EMBL-EBI's Job Dispatcher service, making it quicker and more robust. RNAcentral sequence search using HMMER3 nhmmer has also been added to the Job Dispatcher tools listing, alongside EMBL-EBI's other search and analysis services.

• Species-subsetted FASTA files on the FTP: we now provide separate FASTA files for species with more than 100 sequences in addition to a file containing all active sequences.

• New version of R2DT deployed on the web app: The R2DT web widget has been updated to version 2.0.0, which guarantees the generation of non-overlapping diagrams.

• AI-generated summaries from the API: These are now labelled "WARNING: This summary was generated by AI" when retrieved from the API.

• New preview cards for RNAcentral URLs: When an RNAcentral link is shared (for example on social media), it now displays a meaningful preview card showing the page title and description, instead of a generic logo.

Get in touch

As always, all data are freely available on the RNAcentral website, via the API, our public database, and in the FTP archive. If you have any feedback, please get in touch by email, Bluesky, LinkedIn, or by submitting an issue on GitHub. We look forward to hearing from you!

RNAcentral Release 26

We are thrilled to announce RNAcentral release 26, which introduces a major milestone in how we organize and present non-coding RNA data: the creation of gene-level entries for all sequences across 204 organisms.

Why genes matter

Until now, RNAcentral has been a sequence-based resource, where each unique RNA sequence receives its own identifier (URS id) and is treated as a separate entry. While this approach has served us well, it has created some challenges for our users:

• Fragmented related sequences: Transcripts that differ by just a single nucleotide are treated as completely separate entries, even when they represent variants of the same biological entity.
• Confusing multiplicity: For highly studied genes like rRNAs, there can be thousands of near-identical sequences with slight variations in length or a few nucleotide changes.
• Lost biological context: Many experiments and research questions operate at the gene level rather than the transcript level, making it difficult to find all relevant sequences for a particular gene.

These issues led us to develop a gene building pipeline that groups related transcripts together while maintaining RNAcentral's comprehensive coverage.

Building genes at scale

Creating gene-level entries for RNAcentral presented unique challenges. We couldn't simply adopt gene definitions from resources like Ensembl, because RNAcentral contains many sequences absent from other databases. We needed an automated approach that could:

• Build gene-level entries comparable to those in Ensembl
• Handle the addition of thousands of new sequences with each release
• Maintain stable gene identifiers even as transcripts are added or modified
• Work consistently across all organisms in RNAcentral

Our solution uses a graph clustering algorithm combined with machine learning. We compare pairs of transcripts using a random forest model trained on manually curated human genes from Ensembl/GENCODE. The model considers three types of features: the distance between transcript start sites, the overlap of exons, and the similarity of RNA types. Transcripts predicted to come from the same gene are connected in a graph, and communities within this graph become our genes.

To ensure genes remain stable across releases, we developed logic to track and merge gene clusters between versions. This means that as RNAcentral grows, gene identifiers persist even as new transcript variants are discovered and added.

What's in this release

RNAcentral release 26 now contains 103,814 human non-coding RNA genes built from 600,225 transcripts, covering 56 different RNA types. The most abundant are long non-coding RNAs (lncRNAs) with 65,187 genes, followed by antisense lncRNAs (16,790 genes) and pre-miRNAs (8,560 genes). Beyond human, we have predicted 367,909 genes across 203 species, totalling 1,189,743 transcripts. The average species has ~1800 predicted genes.

The average ncRNA gene in RNAcentral contains 6 transcripts, though this varies widely depending on the RNA type. We've successfully built single genes for well-studied lncRNAs like MALAT1 and NEAT1, and correctly separated miRNA sequences that map to different genomic locations.

Each gene receives a unique identifier following the pattern RNACG<species-prefix><11-digit hash>.<version>, allowing you to track genes across releases. Genes are assigned RNA types and descriptions based on expert database annotations, Rfam families, and R2DT structural information.

How to use genes in RNAcentral

We've made genes accessible throughout the RNAcentral website:

• Text search: Select 'Genes' in the Entry Type facet to see only gene-level results
• Sequence pages: All transcript entries now link to their parent gene (if applicable), e.g. https://rnacentral.org/rna/URS0000D59DC9/9606
• GFF files: Genes appear as 'predicted_gene' entries in our downloadable annotation files, available on our FTP site

We recognize that some complex lncRNAs with extensive alternative splicing may not yet be built perfectly, and we're actively working on improvements. This is the first iteration of our gene building pipeline, and we'll continue to refine it based on user feedback and emerging edge cases.

Database Updates

Release 26 does not update any of the underlying data, so database versions are as they were in Release 25.

Get in touch

As always, all data are freely available on the RNAcentral website, via the API, our public database, and in the FTP archive. If you have any feedback about the new gene-level entries or any other aspect of RNAcentral, please get in touch by email, on Twitter, or by submitting an issue on GitHub. We look forward to hearing from you!

RNAcentral Release 25

We are pleased to announce RNAcentral release 25, which features a major update to the Rfam, update TarBase to v9 and improvements to LitScan.

Welcome Rfam 15.0

Last year, Rfam underwent a major update to release 15.0, you can read more about in this blog post or with our NAR article. This produced a major growth in the number of sequences, with Rfam growing from 1,879,595 to 5,902,604 unique sequences in this release. This release comes with updated genomes used in Rfamseq and updated full sequence data for all Rfam families!

Tarbase updated to version 9

Tarbase produces, curates and delivers high-quality experimentally supported miRNA targets on protein-coding transcripts. In our latest release, Tarbase has been updated to the latest version 9, which includes a six-fold increase in the number of miRNA-gene interactions compared to the previous version. Click here to browse all TarBase sequences.

LitScan updates

Litscan has been updated to scan 13,631,547 ids across 1,273,544 papers. We are testing a new addition to LitScan to filter out false-positive articles that are unrelated to ncRNA. In our evaluations, this significantly reduced the number of articles identified where ncRNA was not the main focus of the work, particularly for identifiers that are common words. We hope you will see an improvement in the articles presented on the literature tab. 

Database updates

These databases were updated in this release to the version stated below:

ENA 2024-02-20 -> 2025-03-06
Ensembl 111 -> 113
Ensembl Fungi 58 -> 60
Ensembl Metazoa 58 -> 60
Ensembl Plants 58 -> 60
Ensembl Protists 58 -> 60
Ensembl/GENCODE human 45/mouse M25 -> human 47/mouse M34
Expression Atlas 2023-10-19 -> 2025-04-01
FlyBase FB2024_01 -> FB2025_01
GeneCards 5.19 -> 5.23
HGNC 2024-02-20 -> 2025-03-06
IntAct 2024-02-20 -> 2024-11-05
MalaCards 5.19 -> 5.23
MirGeneDB 3.0
PDBe 2024-02-20 -> 2024-11-05
PomBase 2024-02-20 -> 2024-11-05
QuickGo 2024-11-05
RefSeq 223 -> 2025-03-06
Rfam 14.10 -> 15.0
SGD 2024-02-20 -> 2025-03-06
SRPDB 2024-02-20 -> 2025-03-06
TarBase v7 -> v9
WormBase 2024-02-20 -> 2025-03-06
ZFIN 2024-02-20 -> 2024-11-05
lncRNAdb 223 -> 223

Sequence Search

Our sequence search tool is still in the process of being updated. While we don’t expect this will cause any problems, there may be some temporary capacity issues. This update should be completed soon, and the service will return to full functionality.

Get in touch

As always, all data are freely available on the RNAcentral website, via the API, our public database, and in the FTP archive. If you have any feedback, please get in touch by email, on Twitter, or by submitting an issue on GitHub. We look forward to hearing from you!

RNAcentral Release 24

We are pleased to announce RNAcentral release 24, which features a major update to the tmRNA Website, improvements to our genome browser and updates to LitScan and LitSumm. LitSumm now generates summaries using GPT4. Read on for the details.

Welcome tmRNA Website 2.0

Recently, the tmRNA Website has undergone two major changes. First, they have overhauled how they identify tmRNA sequences. They have provided a new dataset, which now provides RNAcentral 96,670 sequences, which are extensively annotated with functional features. For example:

Show above is an example of Candidatus Gastranaerophilaceae tmRNA (URS0002856755/3022868)

Secondly, the tmRNA Website is no longer updated and all data is now hosted at RNAcentral. We will continue to host and update this data as normal. You can browse the tmRNA data by searching expert_db:”tmRNA Website”. 

Improved taxonomic search

We have improved our taxonomic search to allow users to more easily find sequences of interest. In our last release we made it possible for users to search for subspecies, but this wasn’t very user friendly. Now there is an option to find all sequences from that taxon or any subspecies right in the taxonomic search box. For example this search can now find all E. coli and subspecies, like E. coli K-12.

Here is an example of searching for: (TAXONOMY:"562" OR lineage_path:"562"). There is now a simple button to push to get taxonomy aware search.

This change applies to all taxonomic levels, so searching for all sequences for bacterial rRNAs is as easy as (TAXONOMY:"2" OR lineage_path:"2") AND so_rna_type_name:"RRNA". Currently, users have to choose to use subspecies searching, but this may change in the future. Try it out and let us know if you have any feedback!

LitScan updates

LitScan is our text mining tool and is used to associate papers with the RNA sequences they mention. We use our comprehensive collection of RNA names to associate sequences with papers. 

A significant increase in the number of articles was observed in this release. The number of papers increased from 915 thousand to 1.1 million, and there are 2.5 million sequences with associated papers. Additionally, LitScan actively monitors retracted articles, promptly eliminating them from the results. Over the last three months, approximately 1,000 articles have been retracted. You can browse all sequences with associated papers here: RNA AND has_lit_scan:"True".

LitSumm updates

LitSumm, our tool to produce gene summaries for ncRNAs, has been updated to use GPT4. As part of the update, we worked with Sam Griffiths-Jones to help validate selected miRNA summaries. Thanks Sam! Overall, the change to GPT4 has brought some impressive improvements. You can read the details in our updated pre-print or browse the summaries. 

We are looking forward to expanding LitSumm to as many RNAs as possible. We also have a prototype API for fetching these summaries and would be interested in feedback on it. If you have any feedback on the summaries or would like to use LitSumm on your site, please get in contact.

Database updates

These databases were updated in this release to the version stated below:

• Ensembl 110 -> 111

• Ensembl Fungi 57 -> 58

• Ensembl Metazoa 57 -> 58

• Ensembl Plants 57 -> 58

• Ensembl Protists 57 -> 58

• Ensembl/GENCODE human 42/mouse M31 -> human 45/mouse M34

• FlyBase FB2023_05 -> FB2024_01

• GeneCards 5.18 -> 5.19

• lncRNAdb 221 -> 223

• MalaCards 5.18 -> 5.19

• Rfam 14.9 -> 14.10

• RefSeq 221 -> 223

The below databases were imported at their latest version as of 2024-02-20

• ENA

• HGNC

• IntAct

• PDBe

• PomBase

• SGD

• SRPDB

• tmRNA Website

• WormBase

• ZFIN
  
  

Website updates

Various minor errors have been rectified on the website, including the previously missing thin lines denoting introns in the RNAcentral genomic coordinates data. Should you encounter any inaccuracies, kindly inform us.

Get in touch

As always, all data are freely available on the RNAcentral website, via the API, our public database, and in the FTP archive.  If you have any feedback, please get in touch by email, on Twitter, or by submitting an issue on GitHub. We look forward to hearing from you!