Skip to main content

Table 4 The detailed predictive ability of chosen QSAR models using RF

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

  Des SE SP F ACC AUC
1A2_Sub 2D 0.72 ± 0.067 0.73 ± 0.072 0.72 ± 0.051 0.72 ± 0.047 0.78 ± 0.045
2C9_Sub 2D 0.75 ± 0.060 0.77 ± 0.063 0.75 ± 0.037 0.76 ± 0.035 0.84 ± 0.038
2C19_Sub 2D 0.76 ± 0.063 0.79 ± 0.067 0.77 ± 0.043 0.77 ± 0.040 0.85 ± 0.038
2D6_Sub 2D 0.76 ± 0.046 0.82 ± 0.044 0.79 ± 0.033 0.79 ± 0.032 0.86 ± 0.029
3A4_Sub 2D 0.75 ± 0.038 0.79 ± 0.035 0.76 ± 0.026 0.77 ± 0.024 0.85 ± 0.020
1A2_In 2D 0.82 ± 0.010 0.87 ± 0.009 0.84 ± 0.007 0.85 ± 0.006 0.93 ± 0.005
2C9_In 2D 0.71 ± 0.017 0.90 ± 0.008 0.74 ± 0.011 0.83 ± 0.007 0.90 ± 0.005
2C9_In_B* 2D 0.83 ± 0.014 0.81 ± 0.013 0.82 ± 0.010 0.82 ± 0.008 0.89 ± 0.007
2C19_In 2D 0.81 ± 0.010 0.84 ± 0.008 0.81 ± 0.007 0.83 ± 0.006 0.89 ± 0.005
2D6_In 2D 0.46 ± 0.016 0.97 ± 0.017 0.59 ± 0.015 0.87 ± 0.011 0.87 ± 0.009
2D6_In_B* 2D 0.74 ± 0.018 0.83 ± 0.018 0.77 ± 0.014 0.79 ± 0.012 0.87 ± 0.011
3A4_In MACCS 0.73 ± 0.015 0.86 ± 0.009 0.76 ± 0.010 0.81 ± 0.008 0.89 ± 0.006
  1. B* refers to the balanced model using the random sampling method