Cell Models & Organoids
Human-relevant models for confident decisions
Better data.
Better cells.
Better outcomes.
The technology behind the GPS.
Capybara™ measures cell identity. CellOracle™ predicts how to move cells toward a target fate. Powered by single-cell transcriptomics, together they map where cells are, where they need to go, and the route to get there.
Signal
scRNA-seq
Measure
Capybara
Predict
CellOracle
Output
Iterative refinement
Building the right cells is still too slow, uncertain, and hard to scale.
Cell-based work depends on reaching the intended biological state. Limited markers and functional readouts can miss maturity, fidelity, and composition gaps. CapyBio makes those gaps measurable, then points to how to close them.
Existing approach
Limited readouts provide an incomplete picture of cell identity.
Suboptimal cells can look “good enough” while important biological gaps remain hidden.
Teams rely on slow trial and error to improve protocols.
CapyBio approach
Quantifies how close cells are to the desired reference state.
Reveals maturity, fidelity, and composition gaps that standard assays may miss.
Guides teams toward targeted changes that produce better cells faster.
The same 60 cells, scored two ways.
QC Readout · Marker Panel
Standard60 / 60 marker pass
PDGFRA+ · ACTA2− · THY1+
Pass rate100%
A standard marker panel (left) reports a clean 100% pass; CapyBio’s identity scoring (right) surfaces hybrid and unresolved cells against the Tabula Sapiens reference menu.
Data foundation · scRNA-seq
Single-cell transcriptomics: the identity foundation.
scRNA-seq profiles thousands of individual cells at once, exposing rare, transitional, and off-target states that bulk assays often hide.
Rare population detection
Identify low-frequency cell types that are diluted or invisible in bulk data.
Transitional state mapping
Capture cells mid-differentiation and trace trajectories as they shift identity.
Heterogeneity exposure
Distinguish incomplete or off-target differentiation within an engineered population.
Readout · Single-Cell Embedding
Cells profiled
10⁴–10⁶
Populations
4
Resolution
Per cell
Why it’s central
scRNA-seq is the foundation of CapyBio’s approach. It maps cell states as engineered cells move toward a target and anchors each benchmark to primary human references.
The result
Resolution, scale, and reference context make scRNA-seq the blueprint for two core applications:
Measuring cell identity
assessing how closely engineered cells match their intended state
Directing optimal differentiation
informing the conditions and checkpoints needed to get them there
Readout · Cell-Type Composition
REF · MCA-NEONATALDiscrete
68.7%
Multi ID
21.4%
Unknown
9.9%
Capybara: scoring how well a cell matches its target.
Gene expression alone does not show how closely an engineered cell matches its target. Capybara benchmarks each cell against primary human references and returns continuous identity scores.
Key Advantages
Quantitative identity confirmation
Score how closely each cell matches its intended target type, with single-cell resolution.
Off-target detection
Identify cells that have adopted unintended identities, even inside the correct cluster.
Population-level purity
Understand the true composition of an engineered batch, not just its dominant cell type.
Protocol benchmarking
Compare differentiation conditions, timepoints, or manufacturing runs against a consistent standard.
Reference-anchored
Scores are grounded in primary human tissue data, not relative comparisons inside your own dataset.
Why it matters
Cluster labels say what a cell resembles. Capybara quantifies how close it is to the target, supporting manufacturing consistency and protocol optimization.
CellOracle: predicting which perturbations drive differentiation.
Identifying cell state is only half the challenge. CellOracle models which transcription factors to perturb to push cells toward the desired identity before running the experiment.
Key Advantages
In silico perturbation screening
Test hundreds of TFs computationally before committing to wet-lab work.
Cell-state-specific modeling
GRNs are built from the actual regulatory logic of each cell state, not a generic network.
Single-cell and population predictions
Understand how perturbations affect individual cell states and the overall trajectory.
Target prioritization
Identify TFs most likely to drive efficient, clean transitions toward the desired cell identity.
Reduced experimental waste
Focus wet-lab resources on the interventions most likely to succeed, shortening timelines.
Perturbation Score · Knockdown
Purple bars: knocking these TFs down pushes cells toward the target identity. Orange bars: knocking these down pushes cells away, meaning they are required for the target fate.
Console · Vector Field
DepleteAccumulate
Predicted direction of cell motion
Predicted cell flow under Ets1 knockout
Select a TF to simulate knockout
Why it matters
CellOracle turns cell engineering from reactive to predictive. Teams enter the lab with a ranked shortlist, reducing failed experiments and shortening discovery.
Applications
Precise human cells, across the things that matter.
Precise human cells for research, pre-clinical testing, therapy, regenerative medicine, and more.
Cell Therapy Manufacturing
Optimal identity with maximal yield
Regenerative Medicine
Better cells, better function
A platform that doesn’t just measure success.
It accelerates it.
Capybara confirms what your cells are and how faithfully they match the target.
CellOracle predicts the next move. Together, they close the loop: measure, predict, engineer, and measure again.