Fig. 5From: MRCNN: a deep learning model for regression of genome-wide DNA methylationClustering results for hypermethylation and hypomethylation loci of the original features and the learned filters of the first convolutional layer. a t-SNE plot of the original input tensor representation. From the plot, we cannot apparently distinguish between hypermethylation and hypomethylation. b t-SNE plot of the learned feature map representation. Hypermethylation and hypomethylation are generally groupedBack to article page