[spora] Spatial Proteomics Dataset

AI ready spatial proteomics dataset

Download and explore standardized multiplexed imaging data to understand tissue architecture across cancer types and organs with spora.

MX Images

12,596

Patients

5,254

Cohorts

Technologies

A project by:

Diverse Data Collection

spora spans 31 cohorts across CRC, TNBC, NSCLC, Lymphoma, Melanoma, PDAC, and more, captured across IMC, CODEX, Orion, and MIBI platforms.

Explore included cohorts ›

Standardized file types and annotations for each sample:

Multiplex Images

Paired H&E Images

Tissue Masks

Cell Masks

Cell Types

Patient Data

Meta Data

Dataset Browser

Browse and filter all 31 cohorts by cancer type, technology, and more. Download any cohort directly from the command line or through the spora-io Python library.

View all cohorts ›

Cohort	MX Imgs	Type	Technology	Cancer
danenberg2022breast Nat Genetics 2022 ↗	718	TC	IMC	Breast
lin2023high Nat Cancer 2023 ↗	41	WSI	Orion	Colorectal
schurch2020coordinated Cell 2020 ↗	251	TC	CODEX	Colorectal
cords2024cancer Cancer Cell 2024 ↗	2,040	TC	IMC	Lung
hoch2022multiplexed Sci Immunol 2022 ↗	50	TC	IMC	Melanoma

spora [io]

Load any single or combination of datasets in two lines of Python. The spora-io library provides a unified interface to browse, tile, and load spatial proteomics data with matched H&E and diverse intensity standardizations.

Read the docs ›

                                
                                
                                
                            
example.py

                            from spora_io import ComposedImagingDataset

# Load H&E + IMC images
dataset = ComposedImagingDataset(
    name="danenberg2022breast",
    path="/path/to/dataset",
    modalities=["he", "imc"],
    resolution=1.0,
    crop_size=224,
    modality_kwargs={"imc": {"standardization": "identity"}},
)

tissue_ids = dataset.get_tissue_ids()
composed = dataset.get_composed_tissue(tissue_ids[0])
composed.modalities["he"]  # H&E Tissue
composed.modalities["imc"] # Multiplex Tissue (C, H, W)
# .channel_names  — marker names
# .uniprot_ids    — aligned UniProt IDs
                        

spora [bench]

Benchmark multiplex foundation models across cell-level and patient-level tasks with standardized evaluation protocols.

View benchmarks ›

Cell Phenotyping Macro F1-score
Dataset	KRONOS	VirTues	astir
cords2023cancer	0.599	0.763	0.414
danenberg2022breast	0.356	0.605	0.358
hoch2022multiplexed	0.576	0.842	0.431
lin2023high	0.393	0.639	0.275
meyer2025stratification	0.480	0.704	0.424
moldoveanu2022spatially	0.532	0.688	0.491

Explore Tissue Data

Explore every cohort in spora through an interactive multiplex visualizer with multi-channel overlays, right in your browser.

Launch Tissue Visualizer ›

Getting Started

Download the full collection or explore a small cohort as a quick demo.

Download via command line

Browse all cohorts and download spora or specific datasets using rclone from your terminal.

Open Dataset Browser ›

Download a small cohort directly

Try spora immediately by downloading allam2022spatially, a small lung cancer IMC cohort (26 images, 340MB), straight from your browser.

Download allam2022spatially ›

Citation

If you use spora in your research, please cite:

@article{spora2026,
            title   = {},
            author  = {},
            journal = {},
            year    = {},
            doi     = {}
          }