Geometric-aware and interpretable deep learning for single-cell batch correction via explicit disentanglement and optimal transport
Chenghan Jiang, Ren Zheng, Yongle Ji, Shubin Cao, Yanhua Fang, Zhe Wang, Ruoyu Wang, Shanshan Liang, Shuai Tao
iDLC corrects single-cell batch effects by explicitly disentangling biological and technical features and aligning distributions with geometric constraints, outperforming existing methods in preserving biological structure.
bioRxiv · Feb 20, 2026 · PreprintBioinformatics