Hi-C is a chromosome conformation capture assay used to study three-dimensional (3D) genome organization. Single-cell Hi-C technologies now enable the examination of 3D chromatin organization in individual cells, although these approaches often suffer from low-coverage libraries and data sparsity. Here, we introduce HiC2Self, a self-supervised framework for denoising Hi-C contact maps that requires only low-coverage data as input. HiC2Self reconstructs key structures such as topologically associating domains (TADs) and significant loops from bulk libraries, including cell-type-specific Hi-C structures, without the generalization challenges faced by supervised models. HiC2Self can also accurately reconstruct significant loops from Micro-C data at 1-kilobase resolution. When applied to single-nucleus methyl-3C data, HiC2Self successfully reconstructs local TAD structures around specific genes at 10-kilobase resolution with as few as 50 cells. Last, HiC2Self enables the examination of single-cell structures at 50-kilobase resolution in individual cells of the same cell type. HiC2Self thus provides a general tool for denoising bulk, pseudobulk, and single-cell 3D contact maps to enable downstream analyses.
Yang et al. (Thu,) studied this question.