Table 4.

Genomic tools/algorithm based on deep learning architecture for epigenomics

Tools	DL model	Application	Input/Output	Website Code Source	References
DeepSEA	CNN	To predict multiple chromatin effects of DNA sequence alterations	N/A	https://github.com/Team-Neptune/DeepSea	[165]
FactorNet	CNN + RNN	For predict the cell-type specific transcriptional binding factors (TF)	BED / BED, gzipped bedgraph file	https://github.com/uci-cbcl/FactorNet	[120]
DeMo (Deep Motif Dashboard)	CNN + RNN	For transcription factor binding site perdition (TFBS) by classification task	FASTA / txt	https://github.com/const-ae/Neural_Network_DNA_Demo	[166]
DeepCpG	CNN + GRU	To predict the methylation states from single-cell data	TSV / TSV	https://github.com/cangermueller/deepcpg	[83]
DeepHistone	CNN	To accurately predict histone modification sites based on sequences and DNase-Seq (experimental) data	txt, CSV / CSV	https://github.com/ucrbioinfo/DeepHistone	[84]
DeepTACT	CNN	To predict 3D chromatin interactions	CSV / CSV	https://github.com/liwenran/DeepTACT	[167]
Basenji	CNN	To predict cell-type-specific epigenetic and transcriptional profiles in large mammalian genomes	FASTA / VCF	https://github.com/calico/basenji	[114]
Deopen	CNN	To predict the chromatin accessibility from DNA sequence/ Downstream analysis also included QTL analysis	BED, hkl /hkl	https://github.com/kimmo1019/Deopen	[31]
DeepFIGV (Deep Functional Interpretation of Genetic Variants)	CNN	To predicts impact on chromatin accessibility and histone modification	FASTA / TSV	http://deepfigv.mssm.edu	[62]