Skip to main content
Fig. 2 | BMC Genomics

Fig. 2

From: Deciphering epigenomic code for cell differentiation using deep learning

Fig. 2

Known and novel TF binding motifs learned in the first convolution layers of the CNN models. a. Overlap of motifs learned in the cell models and histone mark models. b. Number of learned motifs shared by different number of cell models. c. Venn diagram showing the number of learned motifs shared by the cell models. d. Number of learned motifs shared by different number of histone mark models. e. Venn diagram showing the number of learned motifs shared by the histone mark models. f. Examples of learned motifs matching known motifs involved in T cell functions. g, h. Hierarchical two-way clustering of the cells and histone marks, respectively, based on the similarity of the learned motifs profiles in the models using hamming distance and average linkage. The Venn diagrams were drew using INteractiVenn [17]

Back to article page