Dataset Viewer
Auto-converted to Parquet Duplicate
input_ids
list
attention_mask
list
entity_id
int64
augmentation_id
int64
cell_types
list
labels
int64
sample_id
string
tissue
string
disease
string
[[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.4764987826347351,0.0,0.0,0.0,0.0,0.0,0.0,0.0(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
0
0
["alpha_immature","delta","beta_stress I","beta_stress I","alpha_immature","beta_stress I","beta_mtD(...TRUNCATED)
1
GSM4453632_T1D5_HPAP032
pancreas
diabetes
[[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.2397037744522095,0.0,0.0,0.0,0.0,0.0,0.0,0.0(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
0
1
["beta_stress I","beta_stress I","beta_stress I","beta_stress I","delta","beta_stress II","beta_stre(...TRUNCATED)
1
GSM4453632_T1D5_HPAP032
pancreas
diabetes
[[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.8960158824920654,0.0,1.8960158824920654,(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
0
2
["delta","alpha_immature","beta_immature","delta","alpha_immature","beta_stress I","delta","beta_str(...TRUNCATED)
1
GSM4453632_T1D5_HPAP032
pancreas
diabetes
[[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
0
3
["beta_stress I","beta_stress I","beta_stress I","beta_stress I","delta","beta_mtDNA deficient","del(...TRUNCATED)
1
GSM4453632_T1D5_HPAP032
pancreas
diabetes
[[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
0
4
["alpha_immature","beta_stress I","delta","delta","beta_mtDNA deficient","beta_stress I","delta","be(...TRUNCATED)
1
GSM4453632_T1D5_HPAP032
pancreas
diabetes
[[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.901142954826355,1.5381653308868408,0.54881942272(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
1
0
["alpha_immature","delta","beta_mtDNA deficient","delta","beta_stress II","alpha_immature","PP","alp(...TRUNCATED)
1
GSM4453628_T1D1_HPAP020
pancreas
diabetes
[[0.32463371753692627,0.32463371753692627,0.0,0.7657335996627808,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
1
1
["alpha_immature","beta_mtDNA deficient","alpha_immature","alpha_mature","delta","delta","alpha_stre(...TRUNCATED)
1
GSM4453628_T1D1_HPAP020
pancreas
diabetes
[[0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.6308040022850037,0.0,0.0,0.0,0.0,0.0,0.6308040022850037,0.6308040022(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
1
2
["delta","alpha_mature","PP","alpha_immature","alpha_immature","beta_stress I","alpha_immature","bet(...TRUNCATED)
1
GSM4453628_T1D1_HPAP020
pancreas
diabetes
[[0.0,0.0,0.0,0.0,0.0,0.0,0.35084983706474304,0.0,0.0,0.0,0.0,0.35084983706474304,0.0,0.350849837064(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
1
3
["alpha_immature","delta","alpha_immature","alpha_immature","delta","alpha_immature","beta_stress II(...TRUNCATED)
1
GSM4453628_T1D1_HPAP020
pancreas
diabetes
[[1.1538125276565552,0.0,0.43344607949256897,0.0,0.0,0.0,0.0,0.43344607949256897,0.0,0.0,0.0,0.0,0.0(...TRUNCATED)
[true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true,true(...TRUNCATED)
1
4
["alpha_mature","alpha_mature","delta","PP","alpha_mature","alpha_mature","delta","delta","delta","b(...TRUNCATED)
1
GSM4453628_T1D1_HPAP020
pancreas
diabetes
End of preview. Expand in Data Studio

Virtual Cell — Patient Example Dataset

A minimal sample dataset for verifying the data format and running quick end-to-end checks with ConvergeBio/virtual-cell-patient.

This dataset is not intended for training or evaluation. It contains a small number of patients and is not representative of a real distribution. Metrics produced from this dataset should not be interpreted.

Dataset contents

Derived from a public type 1 diabetes scRNA-seq study (GSE148073). Preprocessed into the model's input format as a minimal working example.

Split Patients Rows
train 8 40
validation 3 15

Each row is one augmented view of a patient (5 augmentations per patient).

Loading

from datasets import load_dataset

ds = load_dataset("ConvergeBio/virtual-cell-patient-example")
train_ds = ds["train"]
val_ds   = ds["validation"]

Schema

Column Shape Type Description
input_ids [500, 18301] float32 Log-normalized gene expression matrix, aligned to gene_names.txt
attention_mask [500] bool Cell mask (all ones for fixed cell count)
labels scalar int Class index
entity_id scalar int Patient identifier — groups augmented views of the same patient
sample_id scalar str Original sample accession ID

Preparing your own dataset

Input format

Each patient is a single .h5ad (AnnData) file:

adata.X      — cell × gene expression matrix (float32, log-normalized)
adata.obs    — cell-level metadata (cell_type optional)
adata.var    — gene metadata (index must be HGNC gene symbols)

Values should be library-size normalized (target sum 10,000) and log1p transformed. The gene axis must be aligned to the 18,301 genes in gene_names.txt (from the model repo) — missing genes are zero-filled, extra genes are dropped.

Quality control (optional)

Recommended filters before building the dataset:

Parameter Default Description
min genes per cell 200 Remove low-complexity cells
max genes per cell 5,000 Remove likely doublets
max mitochondrial % 10% Remove dying cells

Building the HuggingFace dataset

For each patient, randomly sample 500 cells into a [500, 18301] float32 matrix. Repeat this sampling independently multiple times per patient to create augmented views — each view becomes a separate row with the same entity_id.

Augmentation is strongly encouraged. The model aggregates predictions across views at inference time, producing more robust results. A factor of 5 augmentations per patient is a good default; 1 is supported but not recommended.

Assign each patient a unique integer entity_id. All augmented views of the same patient must share the same entity_id.

The final dataset should be saved in HuggingFace Datasets format:

from datasets import DatasetDict
dd = DatasetDict({"train": train_ds, "validation": val_ds})
dd.save_to_disk("my_dataset")
# or push directly:
dd.push_to_hub("my-org/my-dataset")

Citation

If you use this dataset, please cite the original study:

@article{sachs2022singlecell,
  author    = {Fasolino, Maria and others},
  title     = {Single-cell multi-omics analysis of human pancreatic islets reveals
               novel cellular states in type 1 diabetes},
  journal   = {Nature Metabolism},
  year      = {2022},
  doi       = {10.1038/s42255-022-00531-x},
  note      = {GEO accession: GSE148073},
}

If you use the Virtual Cell patient model, please also cite:

@article{convergecell2026,
  author    = {ConvergeBio},
  title     = {ConvergeCELL: An end-to-end platform from patient transcriptomics to therapeutic hypotheses},
  year      = {2026},
  note      = {Preprint available on bioRxiv},
}

License

CC BY 4.0

Downloads last month
6