New article
The evidence is already there: an interview with Danuta Gawel
Why computational biology has become a risk-reduction tool in pharma R&D, and why most programs still aren’t set up to use it.
The biological intelligence behind better-informed decisions
We turn millions of human transcriptomic profiles into clear answers about the biology behind your disease, your drug or your patient.
For pharma & CRO's
Pharma ServicesOur scientists answer your disease-biology questions, from target discovery into the clinic.
For researchers
Mavatar DiscoveryExplore disease networks built from millions of public gene-expression profiles. No code needed.
For cliniciansIn development
Mavatar PrecisionSupports clinicians in choosing the treatment most likely to work for each patient.
Latest news
What is happening at Mavatar
What we do
Disease biology, read from real human data
The same disease is driven by different biology in different patients. Mavatar maps that biology from millions of human tissue and cell samples. Drug developers use it to choose targets with more confidence, researchers to find what the literature hasn't, and clinicians to understand which treatments are most likely to work. Read more about our publications and validations.
- 20+ years of research
- 25+ disease models
- State-of-the-art machine learning
Mavatar Discovery changes how we approach the early stages of a research question. Instead of relying solely on what we can generate in our own cohort, we can immediately explore how our findings relate to the broader biological landscape - across diseases, tissues and datasets.
How it works
From public transcriptomic data to a disease model you can trace
Every Mavatar disease model is built the same way, from data found in the public domain, and every result can be traced to its original source.
- 8M+ transcriptomes available
- 185M single cells
- 45+ pre-computed networks
We collect public gene-expression data from thousands of studies. Every sample is annotated by tissue, disease and technology, then quality-checked and normalized.
DINA, our network framework, combines hundreds of independent datasets for each disease or tissue. Two genes are linked only when they are consistently co-expressed across those datasets. The network shows relationships that hold up across studies, not single experiments or assumptions from the literature. Our breast cancer network, for example, draws on 598 datasets and over 36,000 samples.
Every connection in the network points to the datasets and samples behind it, so scientists can cross-validate any finding. One model then supports many questions, such as disease mechanisms, biomarkers, patient stratification and drug prioritization. It is updated as new data is published.
Evidence
See what scientists find with Mavatar
Researchers and pharma R&D leaders use our disease models to explore mechanisms, prioritize targets and put their own data in a wider biological context.
- Peer-reviewed · Cited in Journal of Virology, 2026Mapped ADAM17 across human lung cell types to place an RSV mechanism in the right cells before lab work (opens in a new tab)
- Case study · Alzheimer's DiseaseRevealed a shared lysosomal and immune program behind five Alzheimer's risk genes (opens in a new tab)
- Case study · Rare lung diseaseIdentified possible repurposing candidates for a rare lung disease with no treatment guidelines, starting from a single gene (opens in a new tab)
Trusted by
Collaborators and partners
Precision medicine for IBD - RISE, Vinnova Funded
Partners: Pfizer, Takeda Pharma, Örebro Universitet, and Lunds Universitet delivering data-driven resources to identify molecular subtypes of IBD, explore predictors of treatment response, and more.
SOBI - IFNγ-driven mechanisms
Through advanced analytics of public transcriptomic data, we aim to contribute to a clearer understanding of IFNγ-driven mechanisms across relevant biological contexts.
Clínica Universidad de Navarra - Acute Myeloid Leukemia
A research collaboration with Bruno Paiva, one of Europe's most cited researchers in hematological neoplasms to validate candidate biomarkers and explore biological relationships in AML.
Lund University - Alzheimer's Disease
Together with Prof. Tomas Deierborg’s group, we explored galectin-3 and microglial activation in Alzheimer’s disease through brain and disease-specific networks.
Latest news
Stay updated
The Scariest Numbers in Drug Development: A Mavatar Science & Halloween Soirée
One in 29: Mavatar Takes On the Scariest Numbers in Drug Development
Computational biology as risk reduction in pharma R&D: an interview with Danuta Gawel
Curated vs Data-Derived Evidence in AI Drug Discovery
Mavatar enters research collaboration with Clínica Universidad de Navarra on Acute Myeloid Leukemia
EHA2026 Congress
Every better decision starts with a question
Whether you’re exploring disease biology, de-risking a drug program or following our clinical work, we’d like to hear what you’re working on.