Nat Cell BiolJClub
The dual-enhanced graph learning framework DePass allows paired data integration in single-cell and spatial multiomics.
Nature Cell Biology · · Journal Article · Open access
Li, Jiang + more
Abstract ↗AI summary
The abstract is read at the publisher; the summary is JClub's.
DePass achieves superior integration accuracy across 6 modalities, 9 tissue types, and 13 platforms in single-cell and spatial multi-omics data.
- Why it matters: Integrating multimodal data is essential for understanding cellular and tissue complexity but is challenged by higher noise levels and limited applicability of existing methods.
- What they did: DePass employs a coupled enhancement-integration graph learning framework that iteratively denoises data and refines embeddings, tested systematically across diverse datasets.
- The result: DePass not only outperforms other methods in accuracy but also uncovers immune niche substructure and spatial tumor heterogeneity at near single-cell resolution in colorectal cancer data, enabling deeper biological insights.
The findingWhy it mattersWhat they didThe result
- Open access
- 1 cites