Repositioning across species
a therapeutic for one species becomes a candidate for another
One biology.
Cross-species by construction.
Life is interconnected, and our platform treats it that way. Because our models are cross-species by design, every measurement we make improves predictions for humans and animals at once. A discovery in one species can point the way to a cure in another — human or animal alike.
the counterintuitive result
We predict how a drug transfers between species where structure models can't.
Closely related orthologs fold almost identically, so structure-prediction models see no difference between them. Our sequence-based model does: it predicts how binding shifts from one species to another, recovering textbook divergence cases like AHR, CYP11B, and MAO. Spearman ρ ≈ 0.51, where structure-prediction baselines sit near zero.
Novel (sequence-based) ρ ≈ 0.51
Structure-prediction baseline ≈ 0
Cross-species affinity-shift prediction. Higher is better. Method and sample footnoted at publication.
The tree of life, not one branch.
Novel Sites spans 1.34 million proteins across 78 complete proteomes — livestock, companion, primate, and biomedical model species. The same engine reads across all of them.
| Previous largest | Novel Sites | |
|---|---|---|
| Complete proteomes | 10 | 78 |
| Proteins | 101,813 | 1,344,869 |
| Detection methods | 5 | 6 |
Roughly 13× more proteins than the previous largest binding-site dataset (Bender DB). Open, and targeted for release October 2026.
Interested in early access? Get in touch →
what it enables
Caring for people and animals is the same computation.
The companion and farm animals that share our world often suffer for the way they were bred or used. A platform that helps people can help them too — and reading across species turns that into concrete work:
Fewer, better experiments
cross-species prediction reduces animal testing by pointing to the right experiments first
Comparative-oncology bridges
naturally occurring cancers in companion animals, not purpose-bred models, de-risk human relevance
Underserved species
equine, camel, and livestock health that conventional pipelines ignore