Fig. 1From: The impact of cross-docked poses on performance of machine learning classifier for protein–ligand binding pose predictionThe performance in terms of A, D inter-AUROC and intra-AUROC, B, E inter-Rs and intra-Rs, and C, F top1 and top3 success rates based on the random splitting of the PDBbind-ReDocked dataset. For A–C, the crystal poses in the test set is remained while for D–F is removed. The error bars represent the standard deviation of the 10 repetitions, and the dotted lines in C and F indicate the ceiling of the success rateBack to article page