AI-Ready Data Working Group

Home » About » Working Groups » AI-Ready Data Working Group

Objectives

The AIBIO-UK network aims to be the “go-to” place for resources at this interface between AI and the biosciences. To further this aim we would like to create a catalogue of relevant, AI-ready datasets, which we hope will become a useful resource for students and researchers. We will call this catalogue “AIRBDS” (AI-ready Bioscience Datasets), or simply “the catalogue”.  

What constitutes “AI-readiness” in datasets is an active area of research. In order to make AIRBDS as useful as possible, we have established a working group to answer the following headline questions:    

  • What should be the criteria for inclusion in the catalogue?  
  • How should datasets be described?  
  • Where and how should the catalogue be hosted?  
  • What should be the process for compiling the catalogue?  
  • How should the catalogue be maintained?  
  • How can we ensure that the catalogue is accessible and useable? 

AIRBDS Core metric

The AIRBDS Core metric assesses the AI-readiness of datasets. It is designed to be:

  • quantifiable, producing a score and a grade
  • objective, with clear, repeatable scoring criteria
  • simple, comprising 25 yes-or-no questions
  • domain-agnostic, applicable to any type of data
  • application-agnostic, not prescribing a fixed definition of AI
  • evidence-based, allowing researchers to make informed judgements based on machine-readable metadata

Read the preprint on Zenodo here.

Our Github repo contains the metric in machine-readable form, as well as a draft agentic skill for applying it to datasets.

Resources

Here we list some resources that are broadly similar or overlapping in intent to AIRBDS – that is, they collate data sources and expose some information about them.

  • ELIXIR Core Data Resources 
    • Curated set of high-quality data resources – note distinction between “datasets” (e.g. results of a single experiment) and “resources” (queryable collections of data) 
    • Resources are assessed based on five categories: 
      • Scientific focus and quality of science 
      • Community served by the resource 
      • Quality of service 
      • Legal and funding infrastructure, and governance 
      • Impact and translational stories 
    • Methods paper for compiling the list of CDRs: https://f1000research.com/articles/5-2422/v2 
  • GIGAdb 
    • Data repository that stores datasets linked to publications, organised into “data types” which roughly map to subject areas
  • FAIRSharing

Members

Charlie Harrison

Charlie Harrison

Research Software Engineer, AIRBDS lead; University of Aberystwyth

Charlie has worked with startups, governments, charities, universities and multinational corporations as a data scientist, technical consultant and researcher. He has a PhD in bioinformatics, where he examined the role of rare codons in protein folding. He completed a post doc at the University of Bristol on whole-cell models of Escherichia coli. He was previously the Technical Lead on the GSMA’s AI for Impact project, looking at the application of mobile network data for international and humanitarian development, where he worked on standards for sharing mobile location data anonymously, scoped and led implementations using mobile data for transport planning, and contributed to industry-wide ethical guidelines for responsible AI. At AIBIO he is the resident research software engineer, based at Aberystwyth University. He chairs the AI-ready Bioscience Datasets (AIRBDS) working group and sits on the RRI and Ethics working group.

Reyer Zwiggelaar

Reyer Zwiggelaar

AIBIO Co-Director, Ethics/RRI co-lead; University of Aberystwyth

Reyer is the Head of the Graduate School at the Aberystwyth University and has a chair in Computer Science. Reyer has 20+ years collaborative research experience with clinical/biology experts covering the development of machine learning and medical/biological image/data analysis techniques. More recently, he has also been working at the interface with digital repositories and represents Aberystwyth University at Supercomputing Wales. Current funding includes UKRI Centre for Doctoral Training in Artificial Intelligence, Machine Learning and Advanced Computing (EP/S023992/1), Miscanthus AI- Plant Selection and Breeding for Net Zero (EP/Y005430/1).

Annalisa Occhipinti

Annalisa Occhipinti

Teesside University

Annalisa is an Associate Professor at Teesside University, where she works at the intersection between AI and Biology. Her current research topics involve the integration of metabolic modelling and multi-modal deep learning techniques for biomedical and bioprocessing applications.
Annalisa has recently received several awards for her outstanding contributions in the field of machine learning, big data, and cancer research, including an award from the Italian Embassy and the University of Cambridge.
Annalisa is currently leading various research projects funded by BBSRC and EPSRC, developing AI models for healthcare applications and bioprocessing optimisation in collaboration with industrial partners.

Rachel Rusholme Pilcher

Rachel Rusholme Pilcher

Earlham Institute

Rachel is a Senior Postdoctoral Bioinformatician at the Earlham Institute. Her research focuses on understanding crop genetic variation and the identification of pangenomic resources that can be used to develop climate resilient crops for more sustainable agriculture. Rachel works primarily on bread wheat, adapting existing bioinformatic tools and pipelines to work with large complex crop datasets. With a background in transcriptomics, she also uses network analysis to understand how genes work in concert to regulate key processes such as abiotic stress responses. Rachel has a background in crop science and plant molecular genetics, and has over 20 years’ experience working in both academic and industry-led research environments, overseas and in the UK. She is an advocate of collaborative, open and inclusive research cultures, a Women in Wheat champion and a member of the Women in Crop Science initiative. Rachel is a member of the AI-ready Bioscience Datasets (AIRBDS) and EDI (Equality, Diversity and Inclusivity) working groups.

Alexandra Gibbs

Alexandra Gibbs

University of Nottingham

Alex is an Assistant Professor of Agriculture and the Environment at the University of Nottingham. Her research targets plant physiology; understanding the interactions between plant form and function in relation to dynamic growth environments. She combines cross disciplinary methodologies including plant physiology, gas exchange, 3-dimensional reconstruction, empirical modelling and deep learning to link plant canopy structure to the impact on photosynthesis and other metabolic processes. Alex is currently leading a BBSRC International partnership award in AI for biosciences to analyse cross scale water use in plants.

Matt Spick

Matt Spick

University of Surrey

Matt is a lecturer in health and biomedical data analytics at the University of Surrey. His current research interests are focused on two areas. The first relates to the robustness and reproducibility of biomarker research, especially compliance with FAIR Guiding Principles. The second relates to the challenges to the integrity of scientific literature posed by paper mills, including conflicts of interest with the generation and distribution of scientific knowledge, and how policymakers should respond.

Ian Overton

Ian Overton

Queen's University Belfast

Ian Overton FRSB FHEA is a Reader (Associate Professor) of bioinformatics at Queen’s University Belfast where he leads the ‘Data Intensive Biomedicine’ research programme in the Patrick G Johnston Centre for Cancer Research and is a Heath Data Research UK associate. He has long standing interests in machine learning and biomedical data, for example chairing the open data working group at the Royal Society of Edinburgh’s Young Academy of Scotland (2011-2017). Ian’s research includes systems approaches to connect genotype with phenotype in health and disease, integrative network modelling of cell phenotypic plasticity, drug discovery, structural bioinformatics, digital health, and resource/methods development; for example SynLeGG and eHDPrep. He read biochemistry at Oxford University and completed his PhD in bioinformatics in the Hubbard group, University of Manchester working in functional genomics. His postdoctoral research was in the Barton group, University of Dundee and part of the Structural Proteomics of Rational Targets consortium. Ian started his research group in 2009, supported by two consecutive independent research fellowships at the MRC Human Genetics Unit and University of Edinburgh leading an ‘Integrative Network Biology’ programme. His work spans fundamental and translational research, including partnerships with industry.

Justin Clark-Casey

Justin Clark-Casey

BioFAIR

Justin is the lead architect of BioFAIR, coordinating teams that build facilities for workflow creation and data sharing. He has a strong interest in the intersection of AI and digital infrastructure, ranging from model training and system reliability to the implications of AI “co-scientists” in research. Previously, he served as a programme manager at EMBL-EBI, a technical lead on the Human Cell Atlas, and a developer in areas such as blockchain and open-source virtual worlds.
Melanie Vollmar

Melanie Vollmar

EMBL-EBI

Melanie Vollmar is an ARISE Fellow at EMBL-EBI leading a text mining and NLP project for protein structure annotation enrichment. Previously she held Postdoc positions at the Structural Genomics Consortium within the University of Oxford, at the University of Manchester and at Diamond Light Source. She has a degree in biochemistry from the University of Regensburg and a PhD in structural biology from the University of Duesseldorf. Melanie is also a co-editor for Acta Crystallographica Section D: Structural Biology focusing on submissions concerned with AI, computing and statistics. Throughout her career she gained and combined experience in wet lab applications, computational biology, software development, machine learning and AI.
Gavin Farrell

Gavin Farrell

University of Padova; EMBL-EBI

Gavin Farrell is a Computer Science PhD Candidate at the University of Padova and a Visiting Researcher within the Literature Services team at EMBL-EBI. His primary research specialises in artificial intelligence best practices within the life sciences. He aims to improve the transparency of AI methods in scientific literature to support enhanced model and dataset reuse, alongside reproducibility, across the research community. A key focus of his current work is the DOME Registry, a community-driven data resource built to catalogue, assess, and provide standardised AI/ML methods reports based on publications in biology. Additionally, Gavin authored the Open and Sustainable AI (OSAI) guidelines, published in Nature Methods in 2026. This publication outlines an actionable framework to ensure AI models in the life sciences are developed transparently and sustainably supporting long-term reuse, and reproducibility. Before his PhD, Gavin contributed to ELIXIR Europe, working extensively on the Core Data Resource (CDR) accreditation portfolio. He remains heavily involved in the broader European open science landscape through active membership in various European Open Science Cloud (EOSC) projects and working groups, including FIDELIS, EVERSE, and ARENA. His collaborative work within these groups focuses on optimising data resources, ensuring research software quality, and establishing robust AI infrastructures for a thriving open science ecosystem.
Tulsi Suchak

Tulsi Suchak

University of Surrey

Tulsi is a PhD student at the University of Surrey, working on research integrity and FAIR data practices in biomedical science. Her research examines the use and misuse of large health and ‘omics datasets, including the impact of formulaic publications and paper mill activity on scientific quality. Her work aims to strengthen transparency, reliability, and responsible data use in health research.