Skip to main content
. 2024 May 28;16:62. doi: 10.1186/s13321-024-00855-8

Table 3.

Machine learning models used for each target protein within four screening methods, PIC50, pharmacophore, docking, and shape similarity with models’ performance metrics

Target protein Scoring method ML model Model metrics
R2-train R2-val MAE RMSE MSE w_new R2-ext
AA2AR PIC50 SVR RBF 0.910 0.838 0.270 0.339 0.115 0.673 0.891
pharm KNN 0.999 0.880 0.037 0.048 0.002 0.944 0.905
Docking Elastic net Reg 0.887 0.855 0.281 0.339 0.115 0.689 0.864
Similarity Nu-SVR linear 0.943 0.789 0.073 0.089 0.008 0.882 0.818
TDP1 PIC50 Dec. Tree 0.999 0.465 0.125 0.200 0.040 0.549 0.631
pharm Dec. Tree 0.914 0.889 0.030 0.049 0.002 0.955 0.726
Docking Dec. Tree 0.939 0.586 0.241 0.342 0.117 0.510 0.621
Similarity Elastic net Reg 0.755 0.614 0.058 0.077 0.006 0.880 0.667
EGFR PIC50 Random forest 0.790 0.820 0.406 0.434 0.188 0.596 0.797
pharm Adaboost 0.925 0.786 0.009 0.014 0 0.983 0.791
Docking Adaboost 0.992 0.883 0.340 0.403 0.106 0.625 0.721
Similarity Adaboost 0.999 0.863 0.067 0.086 0.007 0.899 0.603
Akt1 PIC50 SVR RBF 0.682 0.743 0.185 0.216 0.047 0.738 0.642
pharm KNN 0.999 0.940 0.022 0.031 0.001 0.969 0.782
Docking SVR RBF 0.812 0.806 0.176 0.242 0.059 0.770 0.700
Similarity Adaboost 0.974 0.805 0.076 0.076 0.011 0.886 0.761
DPP4 PIC50

Nu-SVR

RBF

 0.883 0.847 0.088 0.104 0.011 0.888 0.728
pharm

Nu-SVR

RBF

 0.994 0.632 0.027 0.034 0.001 0.925 0.776
Docking Adaboost 0.942 0.760 0.166 0.187 0.035 0.752 0.716
Similarity Adaboost 0.979 0.901 0.022 0.028 0.001 0.969 0.842
CDK2 PIC50 Random Forest 0.758 0.604 0.319 0.359 0.129 0.553 0.674
pharm Dec. Tree 0.786 0.809 0.012 0.016 0 0.982 0.678
Docking SVR RBF 0.708 0.644 0.074 0.092 0.008 0.872 0.766
Similarity Dec. Tree 0.708 0.828 0.064 0.099 0.010 0.874 0.737
PPARG PIC50 Adaboost 0.939 0.780 0.358 0.412 0.170 0.570 0.711
pharm Gradient boosting 0.999 0.978 0.022 0.027 0.001 0.974 0.810
Docking Nu-SVR RBF 0.903 0.690 0.285 0.324 0.105 0.591 0.739
Similarity KNN 0.999 0.814 0.038 0.047 0.002 0.935 0.605
P53 PIC50 Dec. Tree 0.714 0.676 0.144 0.159 0.025 0.797 0.624
pharm Adaboost 0.985 0.959 0.012 0.015 0 0.986 0.607
Docking Elastic net Reg 0.993 0.906 0.351 0.401 0.161 0.636 0.651
Similarity Adaboost 0.969 0.691 0.086 0.092 0.009 0.834 0.586