This dataset and associated code are released under the CC-BY-NC-ND 4.0 license and may only be used for non-commercial, academic research purposes with proper attribution. - Any commercial use, sale, or other monetization of the htan-wustl dataset and its derivatives, which include models trained on outputs from the htan-wustl datasets, is prohibited and requires prior approval. - By downloading the dataset, you attest that all information (affiliation, research use) is correct and up-to-date. Downloading the dataset requires prior registration on Hugging Face and agreeing to the terms of use. By downloading this dataset, you agree not to distribute, publish or reproduce a copy of the dataset. If another user within your organization wishes to use the htan-wustl dataset, they must register as an individual user and agree to comply with the terms of use. Users may not attempt to re-identify the deidentified data used to develop the underlying dataset.
This dataset is provided “as-is” without warranties of any kind, express or implied. This dataset has not been reviewed, certified, or approved by any regulatory body, including but not limited to the FDA (U.S.), EMA (Europe), MHRA (UK), or other medical device authorities. Any application of this dataset in healthcare or biomedical settings must comply with relevant regulatory requirements and undergo independent validation. Users assume full responsibility for how they use this dataset and any resulting consequences. The authors, contributors, and distributors disclaim any liability for damages, direct or indirect, resulting from dataset use. Users are responsible for ensuring compliance with data protection regulations (e.g., GDPR, HIPAA) when using it in research that involves patient data.

HESCAPE • PyArrow Format

HESCAPE (H&E + Spatial Contrastive Pretraining Benchmark) is a large-scale benchmark for multimodal learning in spatial transcriptomics.
This repository hosts the PyArrow-formatted Hugging Face datasets for HESCAPE, organized by panel as dataset configs.

Available Configs (Panels)

This dataset repo exposes the following configs:

human-5k-panel
human-breast-panel
human-colon-panel
human-immuno-oncology-panel
human-lung-healthy-panel
human-multi-tissue-panel

Each config corresponds to an independent HESCAPE dataset panel.

Schema

Each dataset entry contains the following columns:

Column	Type	Description
`name`	class_label	Unique identifier for the sample
`image`	image	Image patch
`gexp`	array	Transcriptomic expression based on gene panel
`cell_coords`	array	Coords of the image-gexp pair in tissue
`source`	string	Source of data
`atlas`	string	Label for atlas
`age`	string	Age
`cancer`	bool	Whether cancer or not
`oncotree_code`	string	Oncotree code
`tissue`	class_label	Tissue label
`tumor_grade`	string	Grade of tumor
`gender`	string	Gender
`race`	string	Race
`treatment_type`	string	Treatement type
`therapeutic_agents`	string	Therapeutic agent
`tumor_tissue_type`	string	Tumor tissue type
`assay`	string	Assay used
`preservation_method`	string	Preservation method used
`stain`	string	Stain of histology
`spaceranger`	string	Spaceranger version
`species`	string	Species
`cytassist`	string	Boolean

Usage

Load a specific panel (config):

from datasets import load_dataset

# Example: load the human breast panel
ds = load_dataset(
    "Peng-AI/hescape-pyarrow",
    name="human-breast-panel",
    split="all",
    streaming=True
)
print(ds)

List all configs

from datasets import get_dataset_config_names

get_dataset_config_names("Peng-AI/hescape-pyarrow")

How to cite:

@misc{gindra2025largescalebenchmarkcrossmodallearning,
      title={A Large-Scale Benchmark of Cross-Modal Learning for Histology and Gene Expression in Spatial Transcriptomics}, 
      author={Rushin H. Gindra and Giovanni Palla and Mathias Nguyen and Sophia J. Wagner and Manuel Tran and Fabian J Theis and Dieter Saur and Lorin Crawford and Tingying Peng},
      year={2025},
      eprint={2508.01490},
      archivePrefix={arXiv},
      primaryClass={q-bio.GN},
      url={https://arxiv.org/abs/2508.01490}, 
}

Contact:

Rushin Gindra Helmholtz Munich, Munich ([email protected])
The dataset is distributed under the Attribution-NonCommercial-ShareAlike 4.0 International license (CC BY-NC-SA 4.0 Deed)

Downloads last month: 19