scDART: integrating unmatched scRNA-seq and scATAC-seq data and learning cross-modality relationship simultaneously

Table 1 Number of neurons at each layer of scDART, where n_regions, n_genes, and n_latent refer to number of regions, genes, and latent dimensions respectively

	Input dimension	Layer 1	Layer 2	Output dimensions
Gene activity module	n_regions	1024	512	n_genes
Projection module	n_genes	512	128	n_latent

ISSN: 1474-760X