Datasets

This page summarizes the diversity of tissues and benchmark scenarios represented in spCellEval.

13 curated datasets spanning multiple tissue and platform contexts.

Annotation Levels

Multi-resolution labels enable fine-grained and coarse-grained evaluation.

Balanced evaluation of robustness across biological and technical variability.

Comprehensive overview of all datasets in the benchmark collection.

S.No.	Dataset	Platform	Tissue	#cells	#cell_types	Annotation	GT agreement (%)
1	cHL1_MIBI	MIBI	Lymph node	1,669,853	14	Clustering + Pathologist verification	–
2	cHL2_MIBI	MIBI	Lymph node	230,889	12	Clustering + Pathologist verification	91.4
3	cHL_CODEX	CODEX	Lymph node	145,161	16	Clustering + Pathologist verification	–
4	myeloma_IMC	IMC	Bone marrow	76,000	12	Manual visual gating+XGBoost	–
5	Mouse Heart	Lunaphore	Heart (Mouse)	563,180	19	Clustering	–
6	healthy_bone_marrow	CODEX	Bone marrow	1,200,000	36	Clustering	–
7	feto_maternal	MIBI	Decidua	500,000	26	Clustering	–
8	IMMUcan	IMC	Multiple (Cancer lesions)	457,117	15	Manual	94.1
9	cycif_ovca	Rarecyte	Ovary	2,735,711	13	TRIBUS+Manual	–
10	Breast_Jun0524	CODEX	Breast	7,610	14	Manual	–
11	CRC_Feb2623	CODEX	Colon	2,505	17	Manual	–
12	GVHD_Feb1323	MIBI	GI-tract	1,794	17	Manual	–
13	Melanoma_Oct2522	MIBI	Multiple (Melanoma lesions)	21,204	18	Manual	–

Distribution of cell types across datasets. Select a dataset to view the abundance.

Select Dataset:

Interactive hierarchy of celltype composition in all datasets.

Return to the results page for filtering and sorting detailed metric tables.

Back To Results