Datasets:

SiatBioInf
/

SingleCell-Unseen-Benchmark

Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 2 new columns ({'Cell type', 'Model'}) and 4 missing columns ({'Primary_site/cell type', 'Task', 'Category\t', 'Prediction type'}).

This happened while the csv dataset builder was generating data using

hf://datasets/SiatBioInf/SingleCell-Unseen-Benchmark/benchmark results/by_task/results_neural_cell_identification.csv (at revision c0c7cce7e136747baec670190b9abf74767ae46e)

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1831, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 714, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              Model: string
              Cell population: string
              Cell type: string
              Accuracy: double
              Precision: double
              Recall: double
              F1 Score: double
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 1084
              to
              {'Category\t': Value('string'), 'Task': Value('string'), 'Prediction type': Value('string'), 'Cell population': Value('string'), 'Primary_site/cell type': Value('string'), 'Accuracy': Value('float64'), 'Precision': Value('string'), 'Recall': Value('string'), 'F1 Score': Value('string')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1334, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 911, in stream_convert_to_parquet
                  builder._prepare_split(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1833, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 2 new columns ({'Cell type', 'Model'}) and 4 missing columns ({'Primary_site/cell type', 'Task', 'Category\t', 'Prediction type'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/SiatBioInf/SingleCell-Unseen-Benchmark/benchmark results/by_task/results_neural_cell_identification.csv (at revision c0c7cce7e136747baec670190b9abf74767ae46e)
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Category string	Task string	Prediction type string	Cell population string	Primary_site/cell type string	Accuracy float64	Precision string	Recall string	F1 Score string
Tumor	Tumor cell identification	Binary	Primary tumor cells	Overall	0.008	1	0.008	0.017
Tumor	Tumor cell identification	Binary	Primary tumor cells	Lung	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	Pancreas	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	Colorectal	0.004	1	0.004	0.008
Tumor	Tumor cell identification	Binary	Primary tumor cells	Cervical	0.004	1	0.004	0.008
Tumor	Tumor cell identification	Binary	Primary tumor cells	Kidney	0.008	1	0.008	0.016
Tumor	Tumor cell identification	Binary	Primary tumor cells	Liver	0.005	1	0.005	0.01
Tumor	Tumor cell identification	Binary	Primary tumor cells	Ovarian	0.006	1	0.006	0.012
Tumor	Tumor cell identification	Binary	Primary tumor cells	Esophageal	0.001	1	0.001	0.002
Tumor	Tumor cell identification	Binary	Primary tumor cells	Gastric	0.005	1	0.005	0.01
Tumor	Tumor cell identification	Binary	Primary tumor cells	Breast	0.129	1	0.129	0.229
Tumor	Tumor cell identification	Binary	Primary tumor cells	Head-neck	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	Glioma	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	Lymphoma	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	Melanoma	0.001	1	0.001	0.002
Tumor	Tumor cell identification	Binary	Primary tumor cells	Thyroid	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	DSRCT	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	Myeloma	0	0	0	0
Tumor	Tumor cell identification	Binary	Primary tumor cells	Osteosarcoma	0.001	1	0.001	0.002
Tumor	Tumor cell identification	Binary	Primary tumor cells	Retinoblastoma	0.004	1	0.004	0.008
Tumor	Tumor cell identification	Binary	Primary tumor cells	Prostate	0.002	1	0.002	0.004
Tumor	Tumor cell identification	Binary	Primary tumor cells	Seminoma	0.002	1	0.002	0.003
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Overall	0.002	1	0.002	0.005
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Lung	0	0	0	0
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Pancreas	0	0	0	0
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Colorectal	0.004	1	0.004	0.008
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Liver	0.003	1	0.003	0.006
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Ovarian	0.005	1	0.005	0.01
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Gastric	0.004	1	0.004	0.008
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Breast	0.007	1	0.007	0.014
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Head-neck	0	0	0	0
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Melanoma	0	0	0	0
Tumor	Tumor cell identification	Binary	Metastatic tumor cells	Thyroid	0	0	0	0
Tumor	Tumor cell identification	Binary	Circulating Tumor Cells	Overall	0.036	1	0.036	0.069
Tumor	Tumor cell identification	Binary	Circulating Tumor Cells	Lung	0	0	0	0
Tumor	Tumor cell identification	Binary	Circulating Tumor Cells	Colorectal	0.412	1	0.412	0.583
Tumor	Tumor cell identification	Binary	Circulating Tumor Cells	Liver	0	0	0	0
Tumor	Tumor cell identification	Binary	Circulating Tumor Cells	Breast	0	0	0	0
Tumor	Tumor cell identification	Binary	Circulating Tumor Cells	Melanoma	0.037	1	0.037	0.071
Tumor	Tumor cell identification	Binary	Circulating Tumor Cells	Prostate	0.208	1	0.208	0.344
Tumor	Tumor cell identification	Binary	Normal cells	Overall	0.009	1	0.009	0.018
Tumor	Tumor cell identification	Binary	Normal cells	B cells	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Myeloid	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	T/NK cells	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Endothelial cells	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Epithelial cells	0.001	1	0.001	0.001
Tumor	Tumor cell identification	Binary	Normal cells	Fibroblasts	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Mesenchymal cells	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Germ cells	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Neural stem cells	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Glioblasts	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Neuronal IPCs	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Neuron	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Astrocytes	0.134	1	0.134	0.236
Tumor	Tumor cell identification	Binary	Normal cells	Oligodendrocytes	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	Microglias	0	0	0	0
Tumor	Tumor cell identification	Binary	Normal cells	OPCs	0	0	0	0
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Overall	0.123	0.104 (Macro)	0.117 (Macro)	0.130 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Lung	0.081	0.071 (Macro)	0.081 (Macro)	0.150 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Pancreas	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Colorectal	0.007	0.050 (Macro)	0.007 (Macro)	0.014 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Cervical	0.006	0.053 (Macro)	0.006 (Macro)	0.012 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Kidney	0.006	0.071 (Macro)	0.006 (Macro)	0.012 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Liver	0.006	0.067 (Macro)	0.006 (Macro)	0.012 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Ovarian	0.089	0.077 (Macro)	0.089 (Macro)	0.163 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Esophageal	0.366	0.111 (Macro)	0.366 (Macro)	0.536 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Gastric	0.712	0.167 (Macro)	0.712 (Macro)	0.832 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Breast	0.987	0.500 (Macro)	0.987 (Macro)	0.993 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Head-neck	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Glioma	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Lymphoma	0.002	0.063 (Macro)	0.002 (Macro)	0.004 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Melanoma	0.005	0.053 (Macro)	0.005 (Macro)	0.010 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Thyroid	0.002	0.053 (Macro)	0.002 (Macro)	0.004 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	DSRCT	0.002	0.053 (Macro)	0.002 (Macro)	0.004 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Myeloma	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Osteosarcoma	0.004	0.050 (Macro)	0.004 (Macro)	0.008 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Retinoblastoma	0.013	0.050 (Macro)	0.013 (Macro)	0.026 (Macro)
Tumor	Primary site tracing	Multi-class	Primary tumor cells	Prostate	0.005	0.050 (Macro)	0.005 (Macro)	0.010 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Overall	0.097	0.079 (Macro)	0.094 (Macro)	0.122 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Lung	0.118	0.091 (Macro)	0.118 (Macro)	0.211 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Pancreas	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Colorectal	0.023	0.077 (Macro)	0.023 (Macro)	0.045 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Liver	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Ovarian	0.007	0.125 (Macro)	0.007 (Macro)	0.014 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Gastric	0.386	0.143 (Macro)	0.386 (Macro)	0.557 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Breast	0.378	0.333 (Macro)	0.378 (Macro)	0.549 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Head-neck	0.018	0.143 (Macro)	0.018 (Macro)	0.036 (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Melanoma	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Metastatic tumor cells	Thyroid	0.005	0.056 (Macro)	0.005 (Macro)	0.010 (Macro)
Tumor	Primary site tracing	Multi-class	Circulating Tumor Cells	Overall	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Circulating Tumor Cells	Lung	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Circulating Tumor Cells	Colorectal	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Circulating Tumor Cells	Liver	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Circulating Tumor Cells	Breast	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Circulating Tumor Cells	Melanoma	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Tumor	Primary site tracing	Multi-class	Circulating Tumor Cells	Prostate	0	0.000 (Macro)	0.000 (Macro)	NA (Macro)
Stem	Stem cell identification	Binary	Stem.cell	Overall	1	1	1	1
Stem	Stem cell identification	Binary	Stem.cell	Hematopoietic Stem Cells	1	1	1	1
Stem	Stem cell identification	Binary	Stem.cell	Neural Stem Cells	1	1	1	1
Stem	Stem cell identification	Binary	Stem.cell	Mesenchymal Stem Cells	1	1	1	1

End of preview.

SingleCell-Unseen-Benchmark

Overview

SingleCell-Unseen-Benchmark is a large-scale unseen single-cell transcriptomic benchmark designed to systematically evaluate foundation models on cell identification and cell type tracing tasks.
The benchmark covers tumor, stem, neural, and normal cell populations, with a particular emphasis on unseen data distributions, including rare cell types, cross-dataset generalization, and heterogeneous tumor states.

In addition to curated datasets, this repository provides standardized benchmark results for multiple single-cell foundation models, enabling transparent and reproducible comparison.

Dataset Collection

Tumor Cells

Source: GEO
Cancer types: 21
Samples: 2,225
Cells: 1,645,662
Cell states: Primary tumors, metastases, circulating tumor cells (CTCs)

Stem Cells

Source: CELLxGENE
Datasets: 5
Cells: 325,092
Stem cell types: 4

Neural Cells

Source: CELLxGENE
Datasets: 1
Cells: 423,707
Neural cell types: 6

Normal Cells

Source: CELLxGENE
Datasets: 7
Cells: 1,838,991
Normal cell types: 10

Preprocessing

All genes were mapped to HGNC symbols
Cells with fewer than 200 detected genes were removed
Expression matrices are stored in AnnData (.h5ad) format

Cell Type and Malignancy Annotation Strategy

Tumor cells derived from GEO were re-identified using a consensus workflow:

Lineage-level screening based on CancerSCEM 2.0 marker genes
Malignancy confirmation using inferCNV

CELLxGENE-derived datasets retain their original annotations.

This strategy ensures consistent tumor labeling while minimizing dataset-specific bias.

Downstream Benchmark Tasks

The benchmark evaluates foundation models across multiple biologically meaningful tasks:

Category	Task	Prediction Type
Tumor	Tumor cell identification	Binary
Tumor	Primary site tracing	Multi-class
Stem	Stem cell identification	Binary
Stem	Stem cell subtype classification	Multi-class
Neural	Neural cell identification	Binary
Neural	Neural cell subtype classification	Multi-class

Models take high-dimensional cell embeddings as input and perform prediction using lightweight downstream classifiers, isolating representation quality from classifier complexity.

Benchmark Models

The following single-cell foundation models are evaluated:

Geneformer
scFoundation
scGPT
UCE
scLONG

Evaluation Metrics

Binary classification tasks
- Accuracy
- Precision
- Recall
- F1-score
Multi-class classification tasks
- Accuracy
- Macro-Precision
- Macro-Recall
- Macro-F1

Data Format and Access

Data Files

All datasets are provided in AnnData (.h5ad) format.

Note
.h5ad files are not natively supported by the Hugging Face Dataset Viewer.
Users are expected to download the files and load them locally using standard single-cell analysis tools such as Scanpy or Seurat.

Benchmark Results

In addition to raw datasets, we provide complete benchmark evaluation results under the results/ directory.

Design Rationale

by_model/
Provides a model-centric view, facilitating analysis of how a single model performs across different tasks.
by_task/
Provides a task-centric view, enabling direct comparison of multiple models on the same task.

Both views contain identical information and are provided to improve usability, clarity, and reproducibility.

Intended Use

This benchmark is intended for:

Evaluating generalization and robustness of single-cell foundation models
Studying tumor cell identification and origin tracing under unseen conditions
Benchmarking representation quality across diverse biological contexts

The dataset is not intended for clinical decision-making.

Citation

If you use this dataset or benchmark in your work, please cite:

Contact

For questions, issues, or suggestions, please open an issue on the Hugging Face repository.

Downloads last month: 132