Collective feature selection to identify crucial epistatic variants

Table 1 Parameters used for generating simulated experiment 1 data

Type of Effect (100 and 500 SNPs)	Dataset name	Causal SNP Model (Penetrance: 0.1, 0.5 and 0.9)
Main Effect	1SNP	G1
	2SNP	G1, G2
	3SNP	G1, G2, G3
	4SNP	G1, G2, G3, G4
Interaction Effect	case1_control0	G1 < −>G2
	case1_control1	G1 < −>G2 G99 < −>G100
	case2_control0	G1 < −>G2 G3 < −>G4
	case2_control2	G1 < −>G2 G3 < −>G4 G97 < −>G98 G99 < −>G100

All datasets consisted of 2000 cases and 2000 controls (4000 samples in total). ‘G’ here refers to the SNP ID prefix

ISSN: 1756-0381