Skip to main content
. 2019 Oct 30;15(10):e1007436. doi: 10.1371/journal.pcbi.1007436

Fig 4. Performances of EG prediction tools based on cross-sample validation.

Fig 4

(A) Relative AUROCs for all five methods (B) AUPRs for these methods. The model was trained based on K562 dataset while the prediction was made in GM12878 (see Methods).