An open CAR-T single-cell atlas to enable in-depth characterization and rational engineering of CAR-T products
📄 Read the preprint on bioRxiv
Sergio Cámara-Peña*, Paula Rodríguez-Márquez*, Nuria Planell, María E. Calleja-Cervantes, Lorea Jordana-Urriza, Giacomo Cinnirella, Shlomit Reich-Zeliger, Paula Rodríguez-Otero, Esteban Tamariz, Idoia Ochoa, Nir Yosef, Juan R. Rodríguez-Madoz‡, Felipe Prosper‡, and Mikel Hernaez‡
(*Equal contribution; ‡Correspondence: jrrodriguez@unav.es, fprosper@unav.es, mhernaez@unav.es)
We built a CAR-T cell functional atlas from over one million cells across 13 studies, integrating data from patients and healthy donors.
The atlas captures 11 phenotypes, links infusion product composition with clinical response, and reveals sex- and age-dependent effects, metabolic signatures, and rare ICANS-associated populations.
This open-access resource provides a foundation to understand CAR-T cell function and guide the rational design of next-generation therapies.
The code provided in this repository enables full reproduction of the CAR-T Cell Atlas, from raw data preprocessing to integration, annotation, visualization, and public dissemination through a ShinyCell app and scVI-hub.
Together, these resources ensure full reproducibility and facilitate the extension of the atlas to incorporate future CAR-T datasets.
1_Data_Preprocessing/
└─ Scripts to process individual datasets either from Cell Ranger, Drop-seq or authors' count matrix up to QC-filtered objects (prior to integration).
2_Integration_and_Annotation/
└─ Integration of all datasets with scVI and manual cell type annotation using curated markers.
3_Plotting/
└─ Code used to generate all figures and tables for the manuscript (main and supplementary).
4_New_Data_Integration/
└─ Workflow to incorporate new datasets into the atlas (from data preprocessing to scArches-scANVI model transfer - *Example with Jordana's dataset*).
5_Atlas_Sharing/
└─ Scripts and configurations for atlas distribution resources (ShinyCell app, scVI-hub model).
Publicly available scRNA-seq data and associated metadata from 14 healthy donors and 102 patients with hematological malignancies were integrated, yielding 182 samples encompassing 414,000 CAR⁺ CD3⁺ T cells after quality control.
Distribution of key metadata features across samples, including disease status, time point, CAR construct, clinical response, ICANS grade, sex, and age. Time points are categorized as infusion product (IP), early (<2 weeks), mid (2 weeks–3 months), and late (>3 months). Sex is indicated as male (M) or female (F).
Complete manually annotated CAR-T cell atlas showing 11 phenotypes.
| Resource | Link |
|---|---|
| 🧬 Zenodo (Atlas raw data) | https://doi.org/10.5281/zenodo.17213452 |
| 🧠 scVI-hub pretrained model | https://huggingface.co/sergiocamarap/Functional-cart-atlas-model |
| 💻 Interactive ShinyCell app | https://wholebioinfo.shinyapps.io/shinyatlas/ |
If you use this repository, please cite:
An open CAR-T single-cell atlas to enable in-depth characterization and rational engineering of CAR-T products.
Sergio Camara-Pena, Paula Rodriguez-Marquez, Nuria Planell, Maria E Calleja-Cervantes, Lorea Jordana-Urriza, Giacomo Cinnirella, Shlomit Reich-Zeliger, Paula Rodriguez-Otero, Esteban Tamariz, Idoia Ochoa, Nir Yosef, Juan R Rodriguez-Madoz, Felipe Prosper, Mikel Hernaez
bioRxiv 2025.10.11.681788; doi: https://doi.org/10.1101/2025.10.11.681788
All analyses were conducted using Python (v3.8.10) and R (v4.5.1 / v4.1.3).
Package versions are listed below as referenced in the Online Methods section of the manuscript:
- Seurat v4.3.0.1
- DoubletFinder v2.0.3
- DropletUtils v1.14.2
- SeuratDisk v0.0.0.9020
- dreamlet v1.0.3
- zenith v1.4.2
- clusterProfiler v4.2.2
- AUCell v1.30.1
- ShinyCell v2.1.0
- scvi-tools (scVI) v0.20.3
- scGraph v0.1.2
- Scanpy v1.9.5
- milopy v0.1.1
- scArches v0.6.1
- scProportionTest v0.1.2