scHG Server
About scHG
scHG is a supercell framework with high-order graph learning enables scalable multi-omics analysis.
The method outputs predicted cluster labels for each cell, and optionally evaluates performance if ground truth labels are provided.
If you use this, please cite: Yixiang Huang, Yuan Gan, Xinqi Gong*. scHG: A supercell framework with high-order graph learning enables scalable multi-omics analysis. PLOS Computational Biology, 2026, 22(5): e1013851.
Submitting Jobs
- Two data files are required (e.g., omics 1 expression, omics 2 expression).
- Third file (optional): ground truth labels for evaluation.
- Adjust hyperparameters: alpha, beta, gamma, number of clusters.
Job Output
You will get a zip file containing:
- y_pred.csv – Cluster assignments for each cell (cell index, predicted label).
- y_coar.csv – Cell-to-supercell mapping labels (cell index, supercell label).
- performance.txt – Evaluation metrics (ARI, NMI, ACC) if ground truth provided.
Download Example
The example contains two data files (PBMC 10x) and a label file to help you understand the correct format.
pbmc_10x_X2.csv - Second omics
pbmc_10x_label.csv - Ground truth cell type labels (optional)
Expected output: 8 clusters with performance metrics.
Upload Data Files
Submit your multi-omics datasets for scHG clustering
Upload Status
Run scHG Clustering
Execute clustering analysis on uploaded files