Every patient's biology can teach us something. At every scale, across every disease.

STELA, is a pioneering initiative that profiles patients, spanning from molecules and tissue architecture to clinical outcomes — across diseases, institutions, continents and time —  building the multimodal resource that powers the next generation of pharmaceutical breakthroughs.

Become a partner
Why STELA Exists

Because what we learn from one individual patient across diseases, institutions, and geographies, can change outcomes for the next.

The Missed Opportunity

A breast cancer patient in Paris and a colorectal cancer patient in Singapore may share tissue architecture patterns that predict response to the same class of therapy. Our insights remain incomplete unless we measure both patients deeply, standardize their data, and then build models that can reason across data types, diseases, populations, continents and time.

What STELA Makes Possible

A growing, multi-institutional resource where every patient is profiled across complementary modalities — imaging, spatial biology, molecular assays, clinical outcomes — powering foundation models that can learn to generalize biology. Not just within one cohort, but across the full diversity of human disease.

Today, the diagnostic and biological data generated for a patient is used to inform that patient’s care and then it’s lost to a data archive. The tissue, the imaging, the molecular profiling, the clinical outcomes: each hold information that could improve treatment for future patients with the same disease, or even different diseases sharing underlying biology.
But these signatures are locked in institutional silos, measured in snap shots, single modalities, never integrated at the scale that AI learning requires.
Powering M-Optimus

STELA generates the data. M-Optimus learns the biology. Researchers use both to improve care.

M-Optimus is designed to learn cross-modal relationships, identifying molecular drivers of disease by integrating signals across data types that are typically analyzed in isolation. The result: better biomarkers, deeper understanding of complex drug mechanisms, and insight into why some patients respond and others don't. Tools that go directly to the researchers and clinicians working to improve treatment.
STELA Provides

  • Matched multimodal data across thousands of patients

  • Spatial biology, imaging, molecular profiles, clinical trajectories, all linked per patient

  • Standardized across institutions, continents, and cancer types

  • Expanding to new modalities and disease areas over time

Starting with Spatial Omics, Genomics, H&E Pathology, Clinical Records in oncology. Designed to grow.

M-Optimus Learns

  • Which tissue architectures predict molecular states

  • How cell compositions and spatial organization drive drug response

  • Biomarker signatures that span modalities, invisible to any single assay

  • Biology that generalizes across diseases, not locked to one cohort

The more modalities matched per patient, the richer the biological understanding the model can extract.

Our Role

Bioptimus supports and accelerates the world's most ambitious biomedical research programs — as your AI partner

Every patient who contributes to STELA helps build models that will improve care for patients they will never meet — across diseases, across borders, across generations.
The science is collaborative by design: every partner contributes to and draws from a shared, growing understanding of human biology.
Participating institutions in STELA are not data suppliers.
They are scientific partners pursuing some of the hardest questions in biomedicine: understanding why tumors resist treatment, why the same drug works in one patient and fails in another, and how spatial tissue organization shapes each patient’s disease trajectory.
Bioptimus brings the AI infrastructure — the world models, the data engineering, the computational scale — so that researchers can focus on the science. The data from all partners flows back into world models that serve every researcher at partnered institutions and, ultimately, the patients those partners treat.
Participate

We profile your patients more deeply than any single center could - and the models come back to your bedside.

STELA doesn't just work with archived tissue. It generates new multimodal data on your active patient cohorts - spatial biology, molecular assays, advanced imaging, and builds AI models that come back to your researchers and clinicians. With an architecture that is evolving over time.
Models that help explain treatment resistance, identify biomarkers, and inform therapeutic decisions for the patients in your care today. The outcomes of those decisions, in turn, make the models better for everyone.
You Contribute

Patients

Tissue

Clinical records

pre-existing data

You Receive

Spatial omics profiling

M-Optimus models for your researchers and clinicians

Publications

Scientific Ecosystems

Partnering with World-Leading Research Institutions to Advance Science Together

STELA brings together institutions with complementary strengths, not as data suppliers, but as co-investigators in a shared scientific mission. Each partner shapes the research agenda and benefits from the collective resource.
Technology
10x Genomics

Founding spatial transcriptomics platform partner. Xenium in situ technology for sub-cellular resolution. The initiative is designed to incorporate additional platforms and modalities over time.

Research
Broad Clinical Labs

Data generation infrastructure. Deep expertise in single-cell and spatial multi-omics at population scale.

Clinical
Leading Cancer Centers

Prospective cohorts with treatment and outcomes data across US, Europe, and Asia. Named partners available under NDA.

Scientific Governance

Research priorities, data standards, and expansion strategy guided by an independent board spanning computational biology, clinical oncology, and machine learning.
Meet our SAB
Join STELA as a Partner

The patients in the atlas today are shaping better medicine for patients who haven't been diagnosed yet.

Whether you have cohorts to contribute, drug programs to illuminate, research ambitions to accelerate, or capital to deploy: the founding phase is underway and there is room at the table.