Skip to main content

Table 2 The performance of the top-10 models in the test dataset for the substrate prediction model

From: In silico prediction of UGT-mediated metabolism in drug-like molecules via graph neural network

Model

ACC

SP

SE

AUC

MCC

Descriptor_ET

0.815

0.736

0.891

0.883

0.628

Descriptor_RF

0.811

0.760

0.859

0.889

0.616

MACCS_ET

0.815

0.769

0.859

0.859

0.616

Descriptor_SVM

0.811

0.744

0.875

0.882

0.633

MACCS_RF

0.807

0.727

0.883

0.864

0.619

Descriptor_LR

0.795

0.744

0.844

0.877

0.591

MACCS_NN

0.807

0.760

0.852

0.870

0.615

AttentiveFP_attentivefp

0.823

0.785

0.859

0.877

0.647

MACCS_SVM

0.803

0.744

0.859

0.850

0.608

Morgan_ET_1024

0.807

0.727

0.883

0.878

0.619