CapyBio provides a computational platform that measures and predicts cell identity to improve cell engineering for drug discovery and regenerative medicine. Its Capybara tool scores engineered cells against primary human references using single-cell RNA-seq data, while CellOracle ranks transcription factor perturbations to guide cells toward target fates. The platform is grounded in peer-reviewed methods published in Nature and Cell Stem Cell.
Funding
Funding not disclosed
Founders
Product
Problem
Drug discovery and cell-based therapies rely on cells that often fail to fully recapitulate human biology, leading to poor translational outcomes. Standard quality control metrics for stem cell differentiation—such as morphology and marker gene expression—do not reliably indicate whether cells match their native tissue counterparts, resulting in immature or off-target populations that introduce bias into downstream assays and contribute to the ~90% clinical trial failure rate.
Solution
CapyBio provides a closed-loop platform for cell engineering that measures, predicts, and improves cell identity. The Capybara tool analyzes single-cell RNA-seq data to score each engineered cell against primary human reference datasets, quantifying how closely a population matches its target identity and identifying off-target cells. CellOracle then ranks transcription factor perturbations likely to move cells toward the desired fate, and the wet lab applies these predictions before feeding refined cells back into Capybara for another round of measurement. This iterative approach gives scientists and biopharma teams a quantitative, data-driven way to improve cell quality before committing to expensive downstream experiments.
Target Audience
Primary customers are scientists and biopharma research teams developing stem cell-based therapies, disease models, and pre-clinical drug discovery assays who need to ensure their cells accurately mimic human biology.
Features
- Capybara: computational scoring of engineered cells against primary human references using single-cell RNA-seq data
- CellOracle: in silico prediction and ranking of transcription factor perturbations to guide cells toward target fates
- Closed-loop workflow integrating measurement, prediction, and wet-lab validation for iterative refinement
- Peer-reviewed foundational methods published in Nature, Nature Biotechnology, Cell, Cell Stem Cell, and Stem Cell Reports
- Applicable to iPSC-derived cell types, organoids, and disease models for pre-clinical drug discovery
- Add-on availability for single-cell RNA-seq data input