scBOND: High-fidelity bidirectional translation between single-cell transcriptomes and DNA methylomes with adaptability to paired data scarcity
Authors/Creators
Description
A sophisticated framework for bidirectional cross-modality translation between scRNA-seq and scDNAm profiles with broad biological applicability. We show that scBOND(a) accurately translates data while preserving biologically significant differences between closely related cell types. It also recovers functional and tissue-specific signals in the human brain and reveals stage-specific and cell type-specific transcriptional-epigenetic mechanisms in the oligodendrocyte lineage. We further introduce scBOND-Aug, a powerful enhancement of scBOND that leverages biologically guided data augmentation, achieving remarkable performance and surpassing traditional methods in paired data-limited scenarios.
Files
scBOND.zip
Files
(1.3 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:3e0ea244dc0330596e47df653699e6b3
|
1.3 MB | Preview Download |
Additional details
Dates
- Updated
-
2025-08-20
Software
- Repository URL
- https://github.com/BioX-NKU/scBOND