CellDiffusion: Enhancing scRNA-seq data integration with accurate inference of cell-cell association.

CellDiffusion improves scRNA-seq data integration by accurately inferring cell-cell associations, effectively aligning cells across batches while preserving biological variation.

• Why it matters: Integrating single-cell transcriptomic datasets is difficult due to cell type-specific batch effects and biological heterogeneity, which current methods often fail to address properly.
• What they did: CellDiffusion is an unsupervised, graph-based model that propagates cellular features through a diffusion process and learns a cell-cell attention network to identify corresponding populations across datasets, applying cell-specific correction.
• The result: Validation across datasets, including the complex human endometrium, shows that CellDiffusion enhances data integration, label transfer, clustering, and trajectory inference, enabling more accurate single-cell analysis.