Fig. 1
From: Genome-wide prediction of cis-regulatory regions using supervised deep learning methods
Mean performance and standard deviation of 100 runs using the MLP model on our respectively sampled train-test partitions of eight cell types. a Classification performances of A-E versus A-P. b Classification performances of A-E versus I-E. c Classification performances of A-E versus A-P versus BG. MLP: Multilayer Perception, RF: Random Forest, A-E: Active Enhancer, A-P: Active Promoter, A-X: Active Exon, I-E: Inactive Enhancer, I-P: Inactive Promoter, I-X: Inactive Exon, UK: Unknown or Uncharacterized, BG: I-E+I-P+A-X+I-X+UK