A Robust Rerank Approach for Feature Selection and Its Application to Pooling-Based GWA Studies
Table 2
Set-based analysis of the 8 selected genes under different models of association test.
Gene
NSNP
Allelic model
Dominant model
Recessive model
Additive model
NSIG
ISIG
EMP
NSIG
ISIG
EMP
NSIG
ISIG
EMP
NSIG
ISIG
EMP
Gene 1
11
0
0
1.000
1
1
0.375
0
0
1.000
0
0
1.000
Gene 2
12
1
1
0.443
1
1
0.069
0
0
1.000
1
1
0.415
Gene 3
10
0
0
1.000
0
0
1.000
0
0
1.000
0
0
1.000
Gene 4
9
2
1
0.057
2
1
0.190
1
1
0.135
2
1
0.005*
Gene 5
11
4
2
0.008*
3
2
0.011*
1
1
0.008*
4
2
0.005*
Gene 6
10
1
1
0.007*
2
2
0.129
1
1
0.054
1
1
0.010*
Gene 7
12
1
1
0.107
1
1
0.038*
0
0
1.000
1
1
0.088
Gene 8
4
1
1
0.042*
0
0
1.000
1
1
0.046*
1
1
0.037*
NSNP: number of SNPs in a set; NSIG: total number of SNPs below value with 0.05; ISIG: number of significant SNPs also passing LD criterion on an -squared threshold of 0.5; EMP: empirical set-based value (*EMP < 0.05).