scOncoAnnot
Pan-cancer cell annotation for single-cell RNA-seq
scOncoAnnot is a pan-cancer single-cell annotation model developed by fine-tuning scGPT on tumor-related single-cell RNA sequencing data. It identifies malignant cells and major tumor-microenvironment populations such as epithelial, T, B, myeloid, stromal, and endothelial cells.
HumanscRNA-seqPan-cancerscGPT-based
Run scOncoAnnot
Upload an AnnData file, then start analysis under your account in one step.
Accepted file
.h5ad
Maximum size
2 GB
Gene identifiers
HGNC-approved symbols
Recommended matrix
Raw counts

Prefer a full gene list (not HVG-only). The system prioritizes raw counts when available; otherwise it selects a suitable matrix from X or other layers. Only .h5ad files are accepted.

Drop your .h5ad file here
or
Supported: .h5ad · max 2 GB
Resources
Example dataset
PRAD single-cell dataset (.h5ad)
Gene reference
Supported HGNC gene symbols (.csv)
Example data: Chan et al., Science (2022), DOI: 10.1126/science.abn0478
Model
Submitted tasks
Please log in to run tasks and view your own submission history.
Task IDFile nameModelStatusStartEndDurationOperation
No data available