Table 4.
Model | Training accuracy | Testing accuracy | OR(95%CI) | p-value | CVC* |
---|---|---|---|---|---|
X2 | 0.5215 | 0.5093 | 1.18 (0.93-1.51) | 0.1797 | 8/10 |
X2 – X5 | 0.5319 | 0.5199 | 1.30 (1.01-1.66) | 0.041 | 8/10 |
X2 – X5 – X6 | 0.5353 | 0.5210 | 1.82 (1.28-2.58) | 0.0007 | 7/10 |
X2 – X3 – X5 – X6 | 0.5371 | 0.5267 | 1.75 (1.26-2.44) | 0.0008 | 10/ 10 |
The polymorphisms rs406113, rs974334, rs1052133 and rs2284659 showing the highest statistical significance in the combinatorial set1 + set 2 were chosen for gene-gene interaction analysis. *CVC, cross-validation consistency. Testing accuracy, p-value and CVC significant were highlighted in bold.
X2 = rs406113 (GPX6), X3 = rs974334 (GPX6), X5 = rs1052133 (OGG1), X6 = rs2284659 (SOD3).