TABLE 1.

Results of RF-based models using 10 single features.

Cross-validation	Feature	Dimension	AUROC	Sn	Sp	Acc	Mcc
10-fold cross-validation	AAC	20	0.91	0.85	0.84	0.85	0.69
	TPAAC	21	0.90	0.83	0.78	0.80	0.61
	APAAC	22	0.89	0.83	0.81	0.82	0.64
	ASDC	400	0.88	0.89	0.68	0.79	0.59
	DPC	400	0.86	0.87	0.64	0.76	0.53
	DDE	400	0.83	0.84	0.73	0.78	0.57
	GAAC	5	0.75	0.72	0.66	0.69	0.39
	GDPC	25	0.78	0.75	0.71	0.73	0.46
	SOCNumber	2	0.70	0.66	0.62	0.64	0.28
	QSOrder	42	0.89	0.82	0.82	0.82	0.64
Independent set validation	AAC	20	0.96	0.91	0.89	0.90	0.80
	TPAAC	21	0.94	0.83	0.86	0.84	0.69
	APAAC	22	0.97	0.89	0.91	0.90	0.80
	ASDC	400	0.92	0.89	0.75	0.82	0.65
	CKSAAGP	100	0.87	0.77	0.81	0.79	0.58
	DPC	400	0.89	0.88	0.70	0.79	0.59
	DDE	400	0.90	0.89	0.84	0.87	0.74
	GAAC	5	0.76	0.83	0.64	0.73	0.48
	GDPC	25	0.80	0.73	0.72	0.73	0.45
	SOCNumber	2	0.73	0.59	0.69	0.64	0.28
	QSOrder	42	0.95	0.92	0.84	0.88	0.77

Best performance metrics are shown in bold.