HumanST-46M

Paper: NexuST: A Hierarchical Foundation Model for Spatial Transcriptomics (bioRxiv, 2026).

HumanST-46M provides the pretraining H5AD data used by the NexuST project. This release contains 76 H5AD files, with training and validation data in separate directories.

Files and splits

Split	H5AD files	Bytes	Size (GiB)
train/	72	101,834,973,232	94.84
val/	4	5,138,760,318	4.79
Total	76	106,973,733,550	99.63

Files retain their original platform subdirectories and filenames. Only H5AD data are included; the CellWorld memory-mapped representation, sampling indices, and separate downstream datasets are not part of this release.

Validation files:

• val/xenium/Pulmonary_Fibrosis/pulmonary_fibrosis_xenium.h5ad
• val/merfish/adult_umb5958.h5ad
• val/cosmx/cosmx_liver_normal.h5ad
• val/cosmx/cosmx_liver_cancer.h5ad

Download

from huggingface_hub import snapshot_download

# Downloads train/ and val/ into HumanST-46M/h5ad/ to match the local project layout.
snapshot_download(
    repo_id="Haiping-UoM/HumanST-46M",
    repo_type="dataset",
    allow_patterns=["train/**/*.h5ad", "val/**/*.h5ad"],
    local_dir="HumanST-46M/h5ad",
)

To download only the validation set, use allow_patterns=["val/**/*.h5ad"].

Read an individual file with AnnData:

import anndata as ad

adata = ad.read_h5ad(
    "HumanST-46M/h5ad/val/cosmx/cosmx_liver_normal.h5ad",
    backed="r",
)
print(adata.shape)
adata.file.close()

Related repositories

• NexuST model
• NexuST model code

Documentation status

The file counts and sizes above are measured from this release. The dataset name is retained from the project; an exact cell-count audit is not yet documented here. Source attribution, preprocessing, field descriptions, and license terms will be documented separately once verified.

Citation

If you use NexuST or HumanST-46M in your research, please cite:

@article{liu2026nexust,
  title   = {NexuST: A Hierarchical Foundation Model for Spatial Transcriptomics},
  author  = {Liu, Haiping and Zhao, Qian and Lin, Lijing and Zou, Zhiyong and Cai, Wenhao and Sun, Jingyuan and Zhou, Yuxi and Alvarez, Mauricio A. and Gilmore, Andrew and Rattray, Magnus and Frangi, Alejandro F. and Zhou, Hongpeng},
  journal = {bioRxiv},
  year    = {2026},
  doi     = {10.64898/2026.09.22.753590}
}