Table 1.
AUC | Acc | MCC | ||||||||
---|---|---|---|---|---|---|---|---|---|---|
Train:val:test | Protonation | Coordinate | Protonation | Coordinate | Average | SD | Average | SD | Average | SD |
1:1:1 | none | 2D | 0.930 | 0.006 | 0.967 | 0.007 | 0.821 | 0.035 | ||
1:1:1 | dominate | 2D | 0.904 | 0.011 | 0.926 | 0.048 | 0.668 | 0.131 | ||
1:1:1 | neutralize | 2D | 0.890 | 0.006 | 0.919 | 0.032 | 0.619 | 0.115 | ||
1:1:1 | none | 3D | 0.907 | 0.008 | 0.797 | 0.035 | 0.440 | 0.019 | ||
1:1:1 | dominate | 3D | 0.971 | 0.003 | 0.927 | 0.001 | 0.734 | 0.005 | ||
1:1:1 | neutralize | 3D | 0.924 | 0.007 | 0.969 | 0.003 | 0.827 | 0.017 | ||
1:1:1 | none | CORINA | 0.989 | 0.003 | 0.958 | 0.003 | 0.826 | 0.012 | ||
1:1:1 | dominate | CORINA | 0.996 | 0.002 | 0.982 | 0.005 | 0.914 | 0.021 | ||
1:1:1 | neutralize | CORINA | 0.998 | 0.002 | 0.991 | 0.006 | 0.954 | 0.026 | ||
1:1:1 | neutralize | 3D | neutralize | CORINA | 0.798 | 0.016 | 0.707 | 0.020 | 0.302 | 0.018 |
4:4:1 | none | 2D | 0.923 | 0.024 | 0.959 | 0.029 | 0.798 | 0.107 | ||
4:4:1 | dominate | 2D | 0.906 | 0.013 | 0.894 | 0.069 | 0.609 | 0.139 | ||
4:4:1 | neutralize | 2D | 0.898 | 0.019 | 0.903 | 0.059 | 0.621 | 0.125 | ||
4:4:1 | none | 3D | 0.911 | 0.009 | 0.801 | 0.043 | 0.458 | 0.033 | ||
4:4:1 | dominate | 3D | 0.972 | 0.003 | 0.928 | 0.012 | 0.739 | 0.030 | ||
4:4:1 | neutralize | 3D | 0.927 | 0.011 | 0.971 | 0.002 | 0.839 | 0.010 | ||
4:4:1 | none | CORINA | 0.990 | 0.003 | 0.957 | 0.009 | 0.821 | 0.029 | ||
4:4:1 | dominate | CORINA | 0.997 | 0.001 | 0.985 | 0.003 | 0.927 | 0.015 | ||
4:4:1 | neutralize | CORINA | 0.999 | 0.001 | 0.993 | 0.005 | 0.966 | 0.023 | ||
4:4:1 | neutralize | 3D | neutralize | CORINA | 0.802 | 0.014 | 0.684 | 0.043 | 0.311 | 0.021 |
Parameters (angle: 280, MPS: 100, ZF:100, AT: 23%, BR: 14.5 mÅ, BMD: 0.4 Å, BT: 0.8 Å, LR: 0.0008, BS: 108, GoogleNet), n = 3 or 9 for 1:1:1 or 4:4:1, respectively. Maximum values for AUC, Accuray in test dataset (Acc), and Matthews correlation coefficient (MCC) in each dataset are indicated by bold.