Table 1.

External validation of QSAR models for MIEs based on ChEMBL data. For each MIE predicting QSAR the average number of true positives (TP), false positives (FP), true negatives (TN), and false negatives (FN) were reported. The metrics for evaluating the predictivity of the models were sensitivity (SEN), specificity (SPE), balanced accuracy (BA), Matthew’s correlation coefficient (MCC) and area under the ROC curve (AUC). Performance is the average of metrics obtained over 100 different training-test splits.

MIE	TP	FP	TN	FN	SEN	SPE	BA	MCC	AUC
AChE	555.6	63.0	887.2	68.0	0.89	0.93	0.91	0.83	0.96
AMPAR	13.4	20.0	651.0	1.2	0.92	0.97	0.94	0.60	0.99
CAR	9.0	6.8	668.6	1.2	0.88	0.99	0.94	0.72	0.95
CYP2E1	4.0	35.2	645.2	1.0	0.80	0.95	0.87	0.29	0.90
GABAR	20.0	11.0	649.0	6.0	0.77	0.98	0.88	0.69	0.96
KAR	4.4	17.4	663.0	0.6	0.88	0.97	0.93	0.42	0.97
NADHOX	15.2	4.4	665.4	0.4	0.97	0.99	0.98	0.87	1.00
NIS	11.0	0.6	673.4	0.4	0.97	1.00	0.98	0.96	1.00
NMDAR	50.0	27.8	604.4	3.4	0.94	0.96	0.95	0.75	0.98
PXR	35.8	35.8	601.8	13.0	0.73	0.94	0.84	0.57	0.92
RYR	11.0	0.6	673.6	0.2	0.98	1.00	0.99	0.96	0.99
THRα	60.0	23.4	599.2	2.8	0.96	0.96	0.96	0.81	0.99
THRβ	110.2	37.8	500.0	38.4	0.74	0.93	0.84	0.67	0.93
TTR	14.8	44.0	624.0	3.8	0.80	0.93	0.87	0.42	0.94
VGSC	28.4	12.8	639.2	5.0	0.85	0.98	0.92	0.76	0.97