Integrating single-cell multi-omics data through self-supervised clustering

Yuansong Zeng, Jianing Chen, Zixiang Pan, Weijiang Yu, Yuedong Yang

Abstract: Highlights•We propose scFPN to integrate multi-modality features for fusion-compressed representation.•scFPN fuses layered features through a pyramid network and clustering.•scFPN outperforms baselines on six datasets, improving ARI by 16.15\% and NMI by 3.995\%.