Table 1.

Performance measures shown include information gain (IG) and classification accuracy. The first P-value shown is derived from a standard permutation test although the second comes from the explicit test of epistasis

Model	Functions	IG (TF1)	IG (TF2)	IG (Enhancer)	IG (TF1, TF2, Enhancer)	Accuracy	P-value	P-value
1	ADD, ADD	0.079	0.067	0.081	0.002	0.732	<0.001	0.06
2	ADD, MULT	0.303	0.103	0.102	0	0.822	<0.001	0.043
3	MULT, MULT	0.137	0.134	0.148	0.002	0.817	<0.001	<0.001
4	XOR, XOR	0.006	0.003	0.008	0.033	0.644	<0.001	<0.001
5	MOD2, XOR	0.002	0	0.001	0.046	0.621	<0.001	<0.001
6	BITX, XOR	0.003	0.001	0.001	0.04	0.644	<0.001	<0.001
7	BITX, MOD2	0	0	0.002	0.058	0.636	<0.001	<0.001
8	CHS, BITA	0.146	0.002	0.001	0.16	0.7875	<0.001	<0.001