← Single-cell & foundation models
CellularAI
CellularAI is a single-cell & foundation models tool whose code is published by Illumina on GitHub, first public in 2026. Illumina BioInsight AI Lab algorithm that predicts transcriptome-wide single-cell responses to genetic perturbations in unseen cell types and synthesizes integer-count perturbed cells; winning method of the Arc Institute Virtual Cell Challenge 2026.
| Category | Single-cell & foundation models |
|---|---|
| Developer (GitHub) | Illumina |
| Task | Predict CRISPRi perturbation responses for unseen cellular contexts; output prediction.h5ad with 400 synthetic cells per perturbation by default |
| Interface | Python 3.12 package (pip install -e .), task YAML, h5ad inputs |
| Website | www.illumina.com/science/genomics-research/articles/introducing-the-cellularai-algorithm--predicting-single-cell-res.html |
| GitHub | Illumina/CellularAI |
| Released | 2026-11-19 (source) |
| Code license | Apache-2.0 |
| License note | code Apache-2.0 (GitHub); the training summary bundle derived from the Illumina Billion Cell Atlas is access-controlled and requested from Illumina, so full predictions need that bundle, while the synthetic example runs without it |
| What this allows | Permissive (commercial OK) Use, modify and ship it inside closed-source commercial products; keep the license text and copyright notices. What do these mean? |
| License source | api.github.com/repos/Illumina/CellularAI · checked 2026-11-20 |
| Pricing | Open source |
| Tags | #perturbation-prediction#virtual-cell#perturb-seq#single-cell#crispr |
| Last checked | · links verified |
Can I use CellularAI commercially?
Yes. The recorded license is permissive, so closed-source commercial use is allowed if you keep the notices.
- Code: Apache-2.0
- Source: api.github.com/repos/Illumina/CellularAI · license checked 2026-11-20
What the license types mean · not legal advice.
Specs
| Pretraining corpus | — |
|---|---|
| Architecture | — |
| Tasks | — |
| Species | — |
| GitHub stars | 35 |
| Interface | Python 3.12 package (pip install -e .), task YAML, h5ad inputs |
Compare with…
Pick 1–3 other Single-cell & foundation models tools to compare side by side with CellularAI.
Related
R toolkit for single-cell genomics, integration and multimodal analysis.
Scalable Python toolkit for single-cell gene expression analysis.
Deep generative model in scvi-tools that integrates flow cytometry, mass cytometry and CITE-seq data across panels, with protein imputation, differential expression and annotation.