Table 1.

Values of VIM_Trio and the adjusted importance measures VIM_LD and VIM_Adj for the top five interactions found in the application of trioFS to one of the simulated trio data sets that resulted in the value of VIM_Trio closest to the median of the 100 VIM_Trio values. The interaction term S_3D ⋀ S_7D that specifies disease risk in this simulation is the most important finding. The other interactions also contain this term, and thus, the variable importance measure VIM_Adj that allows for this type of over-fitting is boosted. Since SNPs were not in linkage disequilibrium in this simulation, the values for the importance measures VIM_LD and VIM_Trio are identical

Interaction	VIM_Trio	VIM_LD	VIM_Adj
S_3D ⋀ S_7D	14.23	14.23	22.95
$S_{3 D} \land S_{7 D} \land S_{40 R}^{C}$	1.37	1.37	1.37
$S_{3 D} \land S_{7 D} \land S_{39 R}^{C}$	1.18	1.18	1.18
$S_{3 D} \land S_{7 D} \land S_{8 R}^{C}$	0.95	0.95	0.95
S_1D ⋀ S_3D ⋀ S_7D	0.89	0.89	0.89