THE CLYRA PLATFORM
A shared representation of cancer across species.
Clyra’s foundation models combine complementary human and canine cancer data to learn tumor states linked to clinical trajectories and translate them into testable oncology-development hypotheses.
Discuss a Program
Two data systems. One translational objective.
Human cancer data
Molecular profiles, clinical characteristics, treatment information, pathology, and outcomes provide the direct human context for every translational question.
Shared model space
The model aligns comparable biology while preserving important differences between species, cohorts, and data modalities.
Longitudinal canine cancer data
Naturally occurring cancers provide complementary treatment, progression, recurrence, and survival trajectories on timelines that can reveal additional biological signal.
How the platform works.
Bring together molecular, clinical, treatment, pathology, and outcome information relevant to the development question.
02 — Harmonize
Normalize across cohorts, data modalities, disease contexts, and species.
03 — Align
Learn shared biological representations while accounting for species-specific variation.
04 — Link to outcomes
Connect tumor states with progression, recurrence, survival, response, and resistance.
05 — Translate
Generate patient-subgroup and biomarker hypotheses that can move into human validation.
Core capabilities
Cross-species representation learning
Model shared cancer biology while preserving relevant species and cohort differences.
Multimodal integration
Bring molecular, clinical, pathology, treatment, and outcome signals into a common representation.
Longitudinal outcome modeling
Connect baseline and evolving tumor states with real clinical trajectories.
Interpretable subgroup discovery
Identify molecularly defined patient populations and the biological features that distinguish them.
Bring us one asset and one development question.
Discuss a Program