scTenifoldXct
Cell-cell interaction inference via manifold learning — local run, no data leaves this machine.
1
Dataset
Use bundled example dataset
or
Upload .h5ad
My file is already log-normalised (skip normalize + log1p)
2
Configure run
Cell metadata column
Sender cell type
Receiver cell type
Enrichment test
Nonparametric (null distribution)
Chi-squared
Rebuild gene regulatory networks
— a precomputed network is available for this dataset; uncheck to reuse it and skip straight to training.
Advanced options
Ligand-receptor database filter
None (use all pairs)
pairs (keep L-R pairing)
comb (any L × any R)
Training steps
Learning rate
Embedding dimensions
Alpha
(mean vs. variance weight)
Mu
(correspondence scale)
Distance metric
euclidean
cosine
correlation
p-value cutoff
Chi2 degrees of freedom
CPUs for network build
Seed
(optional)
Scale correspondence block to network magnitude
Nonparametric test: drop zero-correspondence background pairs
Chi2 test: apply Benjamini-Hochberg FDR correction
Run analysis
3
Results
Download CSV