Fig. 2From: DeepGRP: engineering a software tool for predicting genomic repetitive elements using Recurrent Neural Networks with attentionFalse positive rates (FPRs) for DeepGRP and dna-brnn, both for five models applied to hg19/chr1, hg38/chr1 and mm10/chr2. Training and validation of these models was performed on hg19/chr11 and hg19/chr20, respectivelyBack to article page