Cite¶
ALHENA¶
If you use ALHENA, please cite it using the metadata provided in
CITATION.cff. On GitHub, the
same citation metadata is available through Cite this repository.
Models and datasets used by ALHENA¶
ALHENA builds on external models and public single-cell datasets that should be cited separately when they contribute to an analysis.
| Component | What it is |
|---|---|
| UNI2-h | Pathology foundation encoder used as the H&E backbone. Distributed separately by its authors, who ask that Chen et al. (Nat Med, 2024) be cited. |
| scConcept | Single-cell foundation model used as the cell-level teacher during contrastive distillation, presented in Bahrami et al. (bioRxiv, 2025). |
| Novae | Neighborhood-level spatial-transcriptomics teacher used during contrastive distillation, presented in Blampey et al. (Nat Methods, 2025). |
| DISCO | Source of eleven of the fourteen prebuilt single-cell reference atlases, presented in Li et al. (Nucleic Acids Res, 2022). DISCO states: "All data and scripts in DISCO are free for academic research." |
| Tabula Sapiens | Source of the prostate and uterus prebuilt references through CELLxGENE. CC BY 4.0, with attribution to Tabula Sapiens Consortium et al. (Science, 2022). |
| Gene Expression Omnibus, GSE208653 | Source of the cervix prebuilt reference. Deposited by Guo et al. (Clin Transl Med, 2023). No restriction stated by GEO. |
| PanNuke | Public nucleus dataset used by the demo and Tutorial 6, presented in Gamper et al. (ECDP, 2019). CC BY-NC-SA 4.0, non-commercial. |
| BioImage Archive S-BIAD1407 | Colorectal cancer H&E slides used by the use case example. CC BY 4.0. |
| Mendeley Data 9z48yfzr63 | Nucleus coordinates and published cell labels used by the use case example. CC BY 4.0. |
| Trahearn et al. (J Pathol, 2025) | Study that describes the BRAF and MISSONI colorectal cancer cohorts. Source of the microsatellite instability status and the 44.65 µm proximity threshold used by the use case example. CC BY 4.0. |
Figures in this documentation¶
The example figures show a region of the H&E slide of FFPE Human Breast with Custom Add-on Panel, the infiltrating ductal carcinoma sample, by 10x Genomics, under CC BY 4.0. The image is cropped, downsampled and overlaid with ALHENA results.
Licenses¶
The ALHENA software is released under the BSD 3-Clause License.
The ALHENA model weights, comprising the LoRA adapters and the projection heads used at inference, are released under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0). ALHENA requires the UNI2-h weights, which are not redistributed and must be obtained from MahmoodLab/UNI2-h under the license and terms of use of their authors.
The precomputed single-cell reference representations are released under the CC BY-NC 4.0 license. They remain subject to the terms of their source atlases, and those derived from DISCO are provided for academic research only, in accordance with the DISCO terms of use. No expression values, gene names or cell barcodes from the source atlases are redistributed.
The table of models and datasets above gives the source of each prebuilt reference, its terms and the study to credit.
Consult the corresponding source and license information before redistributing model or reference artifacts.
Back to: Home