Table 9. Gal4 variant prediction performance with residual profile vector input attributes.

Method	Input attributes	Se	Sp	PPV	BAR	MCC	AUC
10-fold CV classification:
RF	EP scores	0.93	0.89	0.92	0.91	0.82	0.97
	EP scores + variant ID	0.97	0.95	0.97	0.96	0.92	0.99
SVM	EP scores	0.83	0.69	0.79	0.76	0.53	0.85
	EP scores + variant ID	0.89	0.87	0.90	0.88	0.75	0.93
DT	EP scores	0.91	0.88	0.91	0.89	0.79	0.95
	EP scores + variant ID	0.90	0.88	0.91	0.89	0.78	0.96
NN	EP scores	0.78	0.79	0.84	0.79	0.56	0.82
	EP scores + variant ID	0.81	0.79	0.84	0.80	0.59	0.83
10-fold CV regression:
REPTree	EP scores (r = 0.72)	0.90	0.75	0.84	0.83	0.67	–
	EP scores + variant ID (r = 0.80)	0.94	0.78	0.85	0.86	0.74	–
SVR	EP scores (r = 0.53)	0.86	0.62	0.76	0.74	0.50	–
	EP scores + variant ID (r = 0.72)	0.91	0.77	0.84	0.84	0.69	–
A chain—training/B chain—testing:
RF	EP scores	0.95	0.91	0.94	0.93	0.86	0.98
	EP scores + variant ID	0.99	0.98	0.99	0.98	0.97	1.00
REPTree	EP scores (r = 0.63)	0.78	0.81	0.85	0.79	0.58	–
	EP scores + variant ID (r = 0.63)	0.94	0.56	0.75	0.75	0.56	–
B chain—training/A chain–testing:
RF	EP scores	0.93	0.91	0.93	0.92	0.83	0.98
	EP scores + variant ID	0.98	0.98	0.98	0.98	0.96	1.00
REPTree	EP scores (r = 0.65)	0.75	0.83	0.86	0.79	0.57	–
	EP scores + variant ID (r = 0.60)	0.94	0.54	0.74	0.74	0.54	–