About us
Nature already ran the experiments.
Pangenome AI reads four billion years of natural experiments and turns them into predictions researchers can test.
Every species carries a record of what worked and what never survived. We built models that can read it.
Pangenome AI was founded by Professor James McInerney out of decades of research in evolutionary biology — and one conviction: that four billion years of recombination and selection have already tested more genetic combinations than any laboratory ever will.
Our Large Pangenome Models process millions of genomes to recover that evidence. Applying transformer architectures and attention mechanisms to pangenomes — the complete gene repertoire of a species — lets us predict which genetic combinations work together, and which quietly fail.
Who is behind this
founder · scientific lead
“The rules of genetic compatibility are already written down. They are just written in a language nobody had read at scale.”
Professor James O. McInerney — Founder & Chief Scientific Officer. Twenty-five years in evolutionary genomics and computational biology, with a PhD and DSc and work published across Nature, Science and PNAS. Enquiries reach him directly.
Peer-reviewed papers
Research funding awarded
In evolutionary genomics
Appointments & honours
Past President, Society for Molecular Biology and Evolution
Led the field’s principal international society, with more than 5,000 members worldwide.
Chair in Evolutionary Biology, University of Liverpool
With prior senior appointments at the Universities of Nottingham and Manchester.
Head of the School of Life Sciences, University of Nottingham
Responsible for one of the UK’s largest life sciences schools and a £200M budget.
Fellow, American Academy of Microbiology
Elected to an academy of roughly 2,000 microbiologists worldwide.
Fellow, European Academy of Microbiology
Recognising sustained contribution to microbiological research in Europe.
Visiting Professor, Harvard School of Public Health
Ongoing collaboration with computational biology groups in the United States.
Our mission
Make the rules of genetic compatibility predictable.
Decode nature’s blueprints
Recover the genetic rules encoded in four billion years of evolutionary history.
Cut the cost of getting it wrong
Replace experimental trial-and-error with predictions that can be tested cheaply and early.
Address global challenges
Antimicrobial resistance, sustainable energy, and the next generation of biotechnology products.
Democratise access
Put pangenome-scale analysis within reach of researchers who do not run their own compute.
How a Large Pangenome Model works
live · alignment across strains
01 · Variation
Thousands of genomes per species, aligned so that presence, absence and rearrangement are all first-class signal — not noise around a single reference.
02 · Constraint
Transformer architectures learn which combinations of genes co-occur, and which the historical record has never tolerated.
03 · Prediction
Those constraints become testable predictions: compatibility, emerging resistance, and the real limits of engineering.
Where we are
Built in England’s Northern Powerhouse.
We sit inside the UK biotechnology corridor, with world-class universities and research facilities on the doorstep. Our work is supported by the BBSRC and the Leverhulme Trust, and involves collaborations with Harvard and institutions across Europe and Brazil.