Skip to main content
. 2019 Oct 11;36(5):1429–1438. doi: 10.1093/bioinformatics/btz762

Table 1.

BCrystal outperforms eight other protein crystallization predictors on the balanced test data

Models Accuracy MCC AUC F-score Recall Precision NPV
PPCpred 0.672 0.359 0.754 0.616 0.528 0.740 0.635
fDETECT 0.646 0.355 0.778 0.504 0.36 0.840 0.593
Crysalis I 0.777 0.556 0.865 0.767 0.738 0.799 0.758
Crysalis II 0.804 0.61 0.888 0.796 0.767 0.828 0.784
XtalPred-RF 0.650 0.301 0.710 0.654 0.663 0.645 0.655
TargetCrys 0.627 0.255 0.637 0.637 0.656 0.619 0.593
CrystalP2 0.585 0.177 0.608 0.627 0.700 0.568 0.613
DeepCrystal 0.828 0.658 0.903 0.822 0.795 0.851 0.809
BCrystal 0.954 0.908 0.981 0.954 0.970 0.939 0.969

Note: Best results are highlighted in bold.