Skip to main content

Table 5 The detailed predictive ability of the chosen QSAR models using XGBoost

From: Machine learning to predict metabolic drug interactions related to cytochrome P450 isozymes

  Des SE SP F ACC AUC
1A2_Sub 2D 0.72 ± 0.070 0.69 ± 0.073 0.71 ± 0.046 0.70 ± 0.040 0.77 ± 0.041
2C9_Sub 2D 0.79 ± 0.055 0.73 ± 0.072 0.76 ± 0.042 0.76 ± 0.039 0.83 ± 0.040
2C19_Sub 2D 0.76 ± 0.060 0.74 ± 0.070 0.75 ± 0.048 0.75 ± 0.043 0.82 ± 0.039
2D6_Sub 2D 0.80 ± 0.046 0.79 ± 0.044 0.79 ± 0.034 0.79 ± 0.030 0.86 ± 0.029
3A4_Sub MACCS 0.77 ± 0.034 0.76 ± 0.037 0.77 ± 0.023 0.77 ± 0.021 0.84 ± 0.018
1A2_In 2D 0.84 ± 0.011 0.87 ± 0.009 0.85 ± 0.007 0.86 ± 0.006 0.93 ± 0.004
2C9_In_B* 2D 0.83 ± 0.013 0.80 ± 0.014 0.82 ± 0.009 0.82 ± 0.009 0.89 ± 0.008
2C19_In 2D 0.82 ± 0.011 0.83 ± 0.010 0.81 ± 0.007 0.82 ± 0.006 0.89 ± 0.005
2D6_In_B* 2D 0.78 ± 0.018 0.81 ± 0.018 0.79 ± 0.013 0.80 ± 0.012 0.87 ± 0.009
3A4_In MACCS 0.76 ± 0.011 0.83 ± 0.010 0.77 ± 0.009 0.801 ± 0.007 0.88 ± 0.006