Matt Jeffryes

BioStudies – one package for all the data supporting a study

Ninety-seven percent of biological research identifiers (accession numbers) are published in supplementary files rather than in manuscripts. These files are usually in proprietary formats, such as Excel. This makes them invisible to standard search engines and difficult for machines to process. This ‘long tail of data’ is a critical resource for reproducibility and discovery, yet it remains hidden from researchers and curators alike. Although Findable, Accessible, Interoperable and Reusable (FAIR) standards exist, they have not yet been adopted and/or coordinated. To address this issue, we organised a multi-stakeholder workshop with international experts from data repositories, publishing houses, text mining, and standards organisations, at SciDataCon 2025 in Brisbane. The solutions presented at the workshop, along with subsequent community dialogue, revealed that, while stakeholders agree on the necessary standards, key questions remain about how to implement them. Three barriers emerged: timing (should FAIRification occur during or after peer review?), governance (who enforces standards?), and incentives (who funds what?). Our recommendations are as follows: authors should deposit datasets early; publishers should embed FAIR standards in the peer review process; repositories should harmonise metadata to enable cross-platform discovery; funders should require and support FAIR compliance. This paper argues that implementing FAIR standards requires coordinated action across the research ecosystem with regard to technical standards, governance, incentives, and workflows.

Unprecedented breakthroughs are being made in life science research through the application of artificial intelligence (AI). However, adherence to method reporting guidelines is necessary to support their reusability and reproducibility. The DOME Copilot solution extracts structured reports of AI methods using a large language model to help interpret manuscripts. It is a fast and efficient resource capable of scaling to annotate the corpus of global AI literature, unlocking value and trust in published methods.

Motivation: Biocuration is the integration of biological information databases for the enhancement of research. Curation of these databases is challenged by the exponential growth of scientific data and literature. Integration of machine learning and artificial intelligence methods into biocuration workflows may help address this challenge. We report on the discussions, ideas, and recommendations gathered from a workshop “AI and biodata resources: implications for sustainability and best practices in biocuration” at the 18th Annual International Biocuration Conference 2025. Results: Participants agreed that while AI offers transformative potential for efficiency and expanded curatorial capacity, its integration faces substantial hurdles. Key challenges revolve around data and model quality. Reproducibility issues and a lack of open, domain-specific training datasets further compound these problems. Broader concerns include inconsistent data standards, underdeveloped ontologies, unstructured data, legacy systems, and underfunded teams. Despite these issues, several successful AI applications were identified, including tools for literature summarization and workflow assistance. Participants emphasized the need for a refined model of human-AI collaboration requiring clear data provenance and transparency, new skills, and avoiding over-reliance on AI-generated data. The workshop ultimately recommended concerted efforts in infrastructure development, standardization, training, and quality assurance to guide the community toward effective human-AI collaboration that maintains scientific rigor.

Biological knowledgebases facilitate discovery across the life sciences by structuring experimental findings into human-readable and computable formats. These essential resources are maintained by a small number of professional biocurators worldwide and face combined chronic underfunding and the exponential growth of the literature. In this perspective, we review how artificial intelligence, particularly large language models and agentic systems, can augment literature-curation workflows. Applications include literature recommendation, entity recognition, data extraction, summarization, ontology development, and quality control with emphasis on published use cases at Global Core BioData Resources and ELIXIR Core Data Resources. We identify key challenges, including the scarcity of training data, difficulty in extracting complex relationships, and concerns about error propagation. To address these challenges, we propose a human-in-the-loop framework where generative artificial intelligence approaches accelerate routine tasks while curators provide critical evaluation and domain expertise. We also propose practical recommendations for the community, including the creation of shared benchmark datasets, harmonized evaluation frameworks, and best-practice guidelines for transparent human-in-the-loop AI deployment in biocuration. These synergistic partnerships will be critical to ensure biological rigour, accelerating knowledge integration while maintaining the quality essential for trusted biological resources.

Expression Atlas (https://www.ebi.ac.uk/gxa/home) is EMBL-EBI’s comprehensive knowledgebase for gene and protein expression across tissues, cell types, conditions, and multiple species. Since our last update, Expression Atlas has expanded substantially in both content and functionality, now comprising >4500 studies from 67 species, with increased proteomics coverage and updated Genotype-Tissue Expression (GTEx) tissue profiles. The resource also includes hundreds of single-cell RNA-seq experiments spanning 21 species, among them externally analysed community datasets such as Tabula Sapiens and GTEx single-nucleus profiles, allowing exploration of curated atlases while maintaining their original analytical framework. Key methodological advances include a new marker gene analysis module for bulk baseline experiments, alongside workflow updates that improve reproducibility. Expression Atlas data are integrated into EMBL-EBI resources such as Ensembl, UniProt, and Europe PMC and disseminated through collaboration with model organism communities such as FlyBase and Gramene. The resource also supports translational research through the European Diagnostic Transcriptomic Library and integration with the Open Targets platform. Future directions include modernizing analysis pipelines, enhancing programmatic access, and delivering AI-ready data formats, strengthening Expression Atlas as a findable, accessible, interoperable, and reusable (FAIR) community-driven resource for both fundamental and translational discovery.

Protein family databases are an important tool for biologists trying to dissect the function of proteins. Comparing potential new families to the thousands of existing entries is an important task when operating a protein family database. This comparison helps to understand whether a collection of protein regions forms a novel family or has overlaps with existing families of proteins. In this paper, we describe a method for performing this analysis with an adjustable level of accuracy, depending on the desired speed, enabling interactive comparisons. This method is based upon the MinHash algorithm, which we have further extended to calculate the Jaccard containment rather than the Jaccard index of the original MinHash technique. Testing this method with the Pfam protein family database, we are able to compare potential new families to the over 17,000 existing families in Pfam in less than a second, with little loss in accuracy.

We now have access to the sequences of tens of millions of proteins. These protein sequences are essential for modern molecular biology and computational biology. The vast majority of protein sequences are derived from gene prediction tools and have no experimental supporting evidence for their translation. Despite the increasing accuracy of gene prediction tools there likely exists a large number of spurious protein predictions in the sequence databases. We have developed the Spurio tool to help identify spurious protein predictions in prokaryotes. Spurio searches the query protein sequence against a prokaryotic nucleotide database using tblastn and identifies homologous sequences. The tblastn matches are used to score the query sequence’s likelihood of being a spurious protein prediction using a Gaussian process model. The most informative feature is the appearance of stop codons within the presumed translation of homologous DNA sequences. Benchmarking shows that the Spurio tool is able to distinguish spurious from true proteins. However, transposon proteins are prone to be predicted as spurious because of the frequency of degraded homologs found in the DNA sequence databases. Our initial experiments suggest that less than 1% of the proteins in the UniProtKB sequence database are likely to be spurious and that Spurio is able to identify over 60 times more spurious proteins than the AntiFam resource. The Spurio software and source code is available under an MIT license at the following URL: https://bitbucket.org/bateman-group/spurio

Protein sequence similarity search is one of the most commonly used bioinformatics methods for identifying evolutionarily related proteins. In general, sequences that are evolutionarily related share some degree of similarity, and sequence‐search algorithms use this principle to identify homologs. The requirement for a fast and sensitive sequence search method led to the development of the HMMER software, which in the latest version (v3.1) uses a combination of sophisticated acceleration heuristics and mathematical and computational optimizations to enable the use of profile hidden Markov models (HMMs) for sequence analysis. The HMMER Web server provides a common platform by linking the HMMER algorithms to databases, thereby enabling the search for homologs, as well as providing sequence and functional annotation by linking external databases. This unit describes three basic protocols and two alternate protocols that explain how to use the HMMER Web server using various input formats and user defined parameters. © 2017 by John Wiley & Sons, Inc.