Clyra Bio logo

Clyra Bio

Discuss a Partnership

Cross-species AI foundation models

Cross-species AI foundation models

Learn from cancer across species. Build better human therapies.

Learn from cancer across species. Build better human therapies.

Learn from cancer across species. Build better human therapies.

Cross-species AI foundation models for translational oncology.

Cross-species AI foundation models for translational oncology.

Cross-species AI foundation models for translational oncology.

Clyra Bio is building cross-species AI foundation models that align naturally occurring canine cancers with human molecular and clinical data to model conserved tumor states and improve oncology drug-development decisions.

Starting with survival-linked patient stratification in sarcoma.

Canine cancer is our data advantage.

Human oncology is our market.

Canine cancer data + human cancer data

Molecular · clinical · treatment · longitudinal outcomes

Canine cancer data + human cancer data

Molecular · clinical · treatment · longitudinal outcomes

Cross-species AI foundation model

Shared biological representations across species

Conserved tumor states

Survival-linked subgroups → oncology development decisions

Conserved tumor states

Survival-linked subgroups → oncology development decisions

The data bottleneck

Human oncology has a longitudinal data problem.

Human oncology has a longitudinal data problem.

Human oncology has a longitudinal data problem.

Drug development depends on understanding how tumors evolve, respond, and recur. But human oncology datasets are often slow to assemble, fragmented across institutions, and weakly connected to longitudinal treatment and outcome data. The result is limited biological signal at the moments when drug developers must make critical decisions about patient selection, biomarkers, trial design, and therapeutic strategy.

Slow

Human clinical trajectories can take years to mature.

Slow

Human clinical trajectories can take years to mature.

Fragmented

Molecular, treatment, and outcome data are often separated across datasets and institutions.

Fragmented

Molecular, treatment, and outcome data are often separated across datasets and institutions.

Underpowered

Small and heterogeneous cohorts can obscure clinically meaningful patient subgroups.

Underpowered

Small and heterogeneous cohorts can obscure clinically meaningful patient subgroups.

Why Cross-Species

A real-world cancer system hiding in plain sight.

A real-world cancer system hiding in plain sight.

Dogs naturally develop many of the same cancers seen in humans, within intact immune systems and shared environments. Their cancers generate complementary molecular, treatment, recurrence, progression, and survival data on timelines that can reveal patterns difficult to observe in human cohorts alone. Clyra uses these naturally occurring cancers as an additional window into tumor biology—not as a replacement for human data, but as a new source of translational signal.

More biological signal.

Not more artificial models.

Why Cross-Species

A real-world cancer system hiding in plain sight.

Dogs naturally develop many of the same cancers seen in humans, within intact immune systems and shared environments. Their cancers generate complementary molecular, treatment, recurrence, progression, and survival data on timelines that can reveal patterns difficult to observe in human cohorts alone. Clyra uses these naturally occurring cancers as an additional window into tumor biology—not as a replacement for human data, but as a new source of translational signal.

More biological signal.

Not more artificial models.

Naturally Occurring Disease

Cancers arise spontaneously rather than through artificial experimental induction.

Longitudinal Clinical Trajectories

Treatment, progression, recurrence, and survival can be linked to molecular profiles.

Human-Relevant Biology

Shared tumor biology creates opportunities to identify conserved disease states across species.

The Clyra Platform

One biological language across species.

One biological language across species.

One biological language across species.

Clyra is developing cross-species AI foundation models that learn shared representations from canine and human molecular, clinical, treatment, and outcome data. The platform is designed to distinguish species-specific noise from conserved cancer biology, revealing tumor states that carry translational value across species.

01 · Multimodal Cancer Data

Genomic, transcriptomic, immune, treatment, clinical, and longitudinal outcome signals.

02 · Cross-Species Alignment

AI models align comparable biological patterns across canine and human cancers.

03 · Conserved Tumor States

The platform identifies molecular and microenvironment states shared across species.

04 · Outcome-Linked Subgroups

Tumor states are connected to progression, recurrence, survival, response, and resistance.

05 · Development Intelligence

Outputs support patient stratification, trial design, biomarker strategy, and therapeutic decisions.

01

Cross-Species Representation Learning

Model shared cancer biology while accounting for differences between species and datasets.

01

Cross-Species Representation Learning

Model shared cancer biology while accounting for differences between species and datasets.

02

Longitudinal Outcome Modeling

Connect molecular states with treatment trajectories and clinical outcomes.

02

Longitudinal Outcome Modeling

Connect molecular states with treatment trajectories and clinical outcomes.

03

Translational Decision Models

Convert biological patterns into actionable hypotheses for oncology development.

03

Translational Decision Models

Convert biological patterns into actionable hypotheses for oncology development.

First Product

Survival-linked patient stratification.

Survival-linked patient stratification.

Survival-linked patient stratification.

Clyra’s first product identifies molecularly defined patient subgroups with distinct survival and disease-progression trajectories. It is designed to help oncology drug developers determine which biological subgroups drive clinical outcomes, respond differently to treatment, or require separate development strategies.

Clyra’s first product identifies molecularly defined patient subgroups with distinct survival and disease-progression trajectories. It is designed to help oncology drug developers determine which biological subgroups drive clinical outcomes, respond differently to treatment, or require separate development strategies.

Patient Subgroup Map

Identify biologically distinct patient populations using integrated molecular profiles.

Survival-Linked Stratification

Connect tumor states with progression, recurrence, and survival trajectories.

Biomarker Strategy

Prioritize molecular and microenvironment hypotheses for retrospective or prospective validation.

Clinical-Development Decisions

Support trial enrichment, subgroup analysis, response interpretation, and program strategy.

Sharper biological hypotheses before larger clinical commitments.

Starting Point

Starting with sarcoma. Built for oncology.

Starting with sarcoma. Built for oncology.

Starting with sarcoma. Built for oncology.

Sarcoma is Clyra’s initial proving ground because it combines an urgent need for better patient stratification with some of the strongest evidence for shared cancer biology across dogs and humans. Small cohorts, substantial molecular heterogeneity, and fragmented outcome-linked data make sarcoma an ideal setting to demonstrate how cross-species models can clarify clinical signal. The underlying platform is designed to expand beyond sarcoma into additional cancers where cross-species biology and longitudinal outcomes can improve drug-development decisions.

Sarcoma is Clyra’s initial proving ground because it combines an urgent need for better patient stratification with some of the strongest evidence for shared cancer biology across dogs and humans. Small cohorts, substantial molecular heterogeneity, and fragmented outcome-linked data make sarcoma an ideal setting to demonstrate how cross-species models can clarify clinical signal. The underlying platform is designed to expand beyond sarcoma into additional cancers where cross-species biology and longitudinal outcomes can improve drug-development decisions.

Strong Translational Bridge

Naturally occurring canine sarcomas provide complementary evidence relevant to human disease.

High Unmet Stratification Need

Rare and heterogeneous patient populations make clinical signal difficult to interpret.

Platform Expansion Potential

The same cross-species architecture can extend into additional cancer types and development questions.

Published Translational Precedent

Independent published research demonstrated that a model developed from canine osteosarcoma data identified conserved tumor-microenvironment subtypes that were also present in human osteosarcoma. These cross-species subtypes stratified progression-free survival across independent human datasets, supporting the scientific premise that naturally occurring canine cancers can reveal human-relevant tumor states.

Independent published research — not a Clyra study

Patkar et al. Clinical Cancer Research, 2024

View Publication

Canine osteosarcoma data

Conserved tumor states

Independent human datasets

Progression-free survival stratification

Published Translational Precedent

Independent published research demonstrated that a model developed from canine osteosarcoma data identified conserved tumor-microenvironment subtypes that were also present in human osteosarcoma. These cross-species subtypes stratified progression-free survival across independent human datasets, supporting the scientific premise that naturally occurring canine cancers can reveal human-relevant tumor states.

Independent published research — not a Clyra study

Patkar et al. Clinical Cancer Research, 2024

View Publication

Canine osteosarcoma data

Conserved tumor states

Independent human datasets

Progression-free survival stratification

Platform Roadmap

One foundation. Multiple oncology decisions.

One foundation. Multiple oncology decisions.

One foundation. Multiple oncology decisions.

Patient stratification is Clyra’s first commercial application. As the platform learns from additional cancer types, data modalities, therapies, and outcomes, the same foundation can support a broader set of oncology R&D decisions.

Patient stratification is Clyra’s first commercial application. As the platform learns from additional cancer types, data modalities, therapies, and outcomes, the same foundation can support a broader set of oncology R&D decisions.

Now

Patient Stratification

Identify survival-linked and biologically distinct patient subgroups.

Next

Response and Resistance Modeling

Characterize tumor states associated with treatment benefit, non-response, recurrence, and resistance.

Expansion

Target and Therapeutic Discovery

Use conserved cross-species biology to prioritize pathways, targets, combinations, and development hypotheses.

The first product creates the workflow. Each program expands the data and model advantage.

Translational Stratification Pilot

Start with one asset and one development decision.

Start with one asset and one development decision.

Clyra works with select oncology drug developers through an 8–12 week, asset-specific Translational Stratification Pilot. Each engagement focuses on one therapeutic program, one indication, and one high-value stratification question.

8–12

weeks

Translational Stratification Pilot

Start with one asset and one development decision.

Clyra works with select oncology drug developers through an 8–12 week, asset-specific Translational Stratification Pilot. Each engagement focuses on one therapeutic program, one indication, and one high-value stratification question.

8–12

weeks

01

Integrated Molecular Subgroup Map

Patient subgroups defined by molecular, immune, clinical, and outcome-linked signals.

02

Outcome-Linked Stratification Hypothesis

Subgroup hypotheses associated with progression, recurrence, survival, response, or resistance.

03

Treatment-Biology Interpretation

Asset-specific interpretation of target biology, mechanism of action, tumor microenvironment, and resistance.

04

Trial and Biomarker Strategy

Recommendations for enrichment, subgroup analysis, biomarker development, and clinical interpretation.

05

Validation Roadmap

A plan for retrospective human analysis, independent validation, or prospective study design.

Discuss a Translational Program

Discuss a Translational Program

Contact

Build better oncology decisions from biology across species.

Build better oncology decisions from biology across species.

Clyra is working with select oncology drug developers, translational medicine teams, and scientific collaborators to apply cross-species AI to patient stratification and clinical development.

Cross-species AI foundation models for translational oncology.

Contact

Build better oncology decisions from biology across species.

Clyra is working with select oncology drug developers, translational medicine teams, and scientific collaborators to apply cross-species AI to patient stratification and clinical development.

Cross-species AI foundation models for translational oncology.

Discuss a Partnership

Contact Clyra

Cross-species AI foundation models for translational oncology.

Cross-species AI foundation models for translational oncology.

Discuss a Partnership

Contact Clyra

Cross-species AI foundation models for translational oncology.