---
pretty_name: HESRT — Human-reviewed spatial-omics resource
language:
- en
license: other
license_name: source-specific
license_link: https://huggingface.co/datasets/Biogod/HESRT/blob/main/docs/REUSE.md
size_categories:
- 1K<n<10K
tags:
- spatial-transcriptomics
- spatial-omics
- histology
- multimodal
- anndata
- biology
- biomedical
- provenance
configs:
- config_name: reviewed_catalog
  default: true
  data_files:
  - split: reviewed
    path: releases/2026-09-17/catalog.parquet
- config_name: studies
  data_files:
  - split: studies
    path: releases/2026-09-17/studies.parquet
- config_name: visual_preview
  data_files:
  - split: examples
    path: releases/2026-09-17/visual_preview.parquet
---

# HESRT

**Human-reviewed, accession-level spatial-omics artifacts with tissue images, expression matrices and source provenance.**

| Samples (GSMs) | Studies (GSEs) | Expression observations | Compressed packages |
|---:|---:|---:|---:|
| 4,682 | 530 | 54,902,915 | 421.75 GB |

Each sample is a self-contained, checksummed `tar.zst` package. Browse the catalog, select the accessions you need, then download those samples. The full collection contains approximately 656.57 GB of uncompressed member data; **do not clone the repository to try one sample**.

The counts describe sample artifacts and source-defined observations (spots, cells or bins), not independent donors. Every included package has a recorded human visual-review PASS. Source interpretation and outstanding verification annotations remain available in the catalog and [known issues](docs/KNOWN_ISSUES.md).

## See the data

![Eight real HESRT tissue-image previews, labeled by accession and source technology](assets/overview.png)

These eight examples span human cortex, mouse brain, spinal cord, pancreas, liver, Visium HD, FFPE, CytAssist and Xenium labels. They are selected illustrations, not a representative sample or a new quality assessment. Each thumbnail comes from a released package, with source hashes and display details in [preview provenance](assets/preview_manifest.json). The `visual_preview` Dataset Viewer configuration pairs tissue thumbnails with the package's existing alignment-QC graphic.

[Full preview gallery](docs/PREVIEWS.md) · [Usage guide](docs/USAGE.md) · [Schema](docs/DATA_SCHEMA.md) · [Ontology mapping](docs/ONTOLOGY_MAPPING.md) · [Known issues](docs/KNOWN_ISSUES.md) · [Reuse and citation](docs/REUSE.md)

## Access

HESRT is public. No sign-in, access request or form is needed: the catalog and every sample package can be downloaded directly, in a browser or with `huggingface_hub`. The maintainers receive no information about who downloads the data. The original source-specific data-use terms still apply.

## Start with a 3.2 MB sample

Python 3.11+. Download the current example tools; they resolve the pinned data release `2026-09-17-v1.6` to a single commit, verify checksums, extract to disk and open the matrix in backed mode.

```bash
python -m pip install "huggingface_hub>=0.34,<2"
python -c "from huggingface_hub import snapshot_download; snapshot_download('Biogod/HESRT', repo_type='dataset', revision='main', allow_patterns=['examples/*'], local_dir='hesrt_tools', max_workers=4)"
python -m pip install -r hesrt_tools/examples/requirements.txt
python hesrt_tools/examples/quickstart.py --output hesrt_demo
```

This loads `GSM8915497` / `GSE294669`, verifies every package member, and reports matrix dimensions, coordinate shape and image format. For filtering, complete-study downloads, local verification and evaluation-split examples, see the [usage guide](docs/USAGE.md).

## Browse and filter

```python
from huggingface_hub import HfApi, hf_hub_download
import pandas as pd

revision = HfApi().dataset_info(
    "Biogod/HESRT", revision="2026-09-17-v1.6"
).sha
catalog = pd.read_parquet(hf_hub_download(
    "Biogod/HESRT", "releases/2026-09-17/catalog.parquet",
    repo_type="dataset", revision=revision
))
human = catalog.loc[catalog.organism.eq("Homo sapiens")]
print(human[["gsm_id", "tissue", "technology", "n_obs", "archive_bytes"]].head())
```

The native Viewer has three configurations: **reviewed_catalog** (4,682 rows, 96 columns), **studies** (530 rows), and **visual_preview** (eight examples). These are lightweight indexes and previews, not streamed full-resolution matrices. `reviewed` is an inventory split, not a train/test assignment.

### Composition

| Exact catalog label | Samples |
|---|---:|
| Homo sapiens | 2,829 |
| Mus musculus | 1,651 |
| 10x Genomics Visium | 4,135 |
| Spatial Transcriptomics | 267 |
| 10x Genomics Visium for FFPE | 108 |
| 10x Genomics Visium HD | 61 |
| 10x Genomics Visium CytAssist | 41 |
| 10x Genomics Xenium In Situ | 8 |

Technology and tissue labels are source-derived. Labels can be heterogeneous, and null disease values mean unknown. The counts above use exact catalog labels; related technologies are not silently merged.

## Ontology harmonization

The catalog now includes a conservative machine-readable ontology layer while preserving every original metadata field. Organisms map to NCBI Taxonomy; animal anatomy maps to UBERON; plant anatomy maps to Plant Ontology; disease maps to MONDO. No fuzzy match is accepted automatically.

Current coverage is **4,679/4,682 organism records**, **3,021/4,682 tissue labels**, and **1,893/2,398 non-null disease labels**. A further **105 disease labels** are explicitly classified as control/reference/non-disease states rather than being forced into MONDO. Compound tumor/specimen phrases that do not identify an ontology term with sufficient confidence stay unresolved.

Use `*_ontology_status` and `*_ontology_match_type` to distinguish exact mappings, source-validated aliases, controls, ambiguity and unresolved values. See the [ontology mapping protocol and audit](docs/ONTOLOGY_MAPPING.md).

## Package contents

```text
GSM....tar.zst
  spatial.h5ad               Expression matrix and spatial coordinates
  wsi.tif                    Selected tissue image; inspect actual format
  metadata.json              Source-associated sample metadata
  metadata.normalized.json   Normalized values, raw fields and statuses
  provenance.json            Source and processing provenance
  review.json                Recorded visual-review decision
  alignment_qc.png           Existing QC graphic, when available
  alignment_strict_qc.png    Additional QC graphic, when available
  manifest.json              Per-member sizes and SHA256 hashes
```

Expression/image binaries are unchanged by lossless packaging of the selected standardized artifacts. Release 2026-09-17-v1.5 replaced 12 packages and withdrew 6 after every released value was compared with the GEO deposit; the replaced packages store coordinates differently and have no QC graphic. Release 2026-09-17-v1.6 withdrew 9 more packages whose spots do not sit on the tissue in the image (see [known issues](docs/KNOWN_ISSUES.md)). HESRT does **not** claim that every `X` matrix contains source-verified raw UMI counts. A release-wide numerical audit identified eight samples whose `X` matrices contain clear non-integer values: `GSM7781044` (GSE243225) and `GSM8243007`–`GSM8243013` (GSE266244). These source-defined matrices are retained as deposited/standardized rather than reverse-engineered into synthetic counts. Analyses that require integer raw counts should exclude or source-verify these eight samples. The remaining matrices are integer-typed or numerically integer-like in the audit, but that alone is not proof of raw-count provenance. Coordinate frame, units and expression representation should be interpreted using provenance. The name `wsi.tif` does not guarantee a TIFF file, a specific stain, or a full-resolution whole-slide scan.

## Release files

The immutable release namespace is `releases/2026-09-17/`. [Sample catalog](releases/2026-09-17/catalog.parquet) and [CSV fallback](releases/2026-09-17/catalog.csv) support selective download. [Study index](releases/2026-09-17/studies.parquet), [schema](releases/2026-09-17/schema.json), [metadata coverage](releases/2026-09-17/metadata_coverage.json), [ontology sidecar](releases/2026-09-17/ontology_mapping.parquet), [ontology mapping summary](releases/2026-09-17/ontology_mapping_summary.json), [package manifest](releases/2026-09-17/packages.json), [SHA256SUMS](releases/2026-09-17/SHA256SUMS), and [release record](releases/2026-09-17/release.json) provide machine-readable provenance. Prior repository paths remain intact.

For evaluation, keep studies and shared content together, resolve donor/cohort overlap and freeze the exact revision. The provided split script is an illustrative leakage-aware starting point, not an official benchmark protocol.

## Related resources

HESRT is the human-reviewed H&E-paired collection produced by the earlier SpatialDataAgent pipeline. Its samples are also integrated into [SpatialDataAgent-DB](https://huggingface.co/datasets/Biogod/SpatialDataAgent-DB) (more than 12,000 samples, most HESRT samples among them after deduplication; see its card for the current release), which uses a different per-sample layout and its own access policy. Do not concatenate the two collections without resolving overlap. The [public SpatialDataAgent-DB catalog](https://huggingface.co/datasets/Biogod/SpatialDataAgent-DB-Catalog) and the [spatial-data-agent-db toolkit](https://github.com/Biogod2020/spatial-data-agent-db) support search and download.

## Reuse and citation

Underlying source materials retain their source-specific terms; no blanket redistribution license is asserted. Cite the source GEO accessions/publications and **HESRT release 2026-09-17-v1.6**, recording its resolved Hub commit. See [reuse and citation](docs/REUSE.md) and [release notes](CHANGELOG.md). This resource is intended for research, not clinical decision making.