Exploring the ability of machine learning-based virtual screening models to identify the functional groups responsible for binding

Table 3 Performance of the RF_PLEC_4 model on different datasets

Dataset	Random accuracy	Accuracy	AU-PRC	Balanced accuracy	Balanced AU-PRC
ZINC	0.504	0.948	0.988	N/A	N/A
DUDE-AA2AR	0.93	0.992	0.987	0.942	0.992
DUDE-DRD3	0.972	0.992	0.974	0.904	0.981
DUDE-FA10	0.97	0.992	0.971	0.934	0.993
DUDE-MK14	0.976	0.99	0.994	0.928	0.985
DUDE-VGFR2	0.984	0.992	0.92	0.926	0.987
LIT-ALDH1	0.613	0.948	0.987	0.946	0.988
LIT-FEN1	0.956	0.968	0.839	0.858	0.948
LIT-MAPK1	0.964	0.972	0.672	0.822	0.923
LIT-PKM2	0.944	0.974	0.976	0.916	0.984
LIT-VDR	0.928	0.98	0.931	0.928	0.979

Accuracy denotes the proportion of correctly classified examples, whereas Random Accuracy denotes the accuracy that would have been obtained by assigning all examples the most common label \((=\text {max}(\%\text { actives, }\%\text { inactives)}\)
AU-PRC denotes the area under the Precision-Recall curve
Balanced Accuracy and Balanced AU-PRC denote the respective accuracy and area under the Precision-Recall curve when the model was trained using an equivalent number of actives and inactives

ISSN: 1758-2946