Table 3b.
Comparisons of Results from the Logistic Regression and Proposed Hierarchical Modeling Approaches – G × E Interactions
| SNP (Gene) | Loaistic Regression | HM1a | HM1b | HM2a | HM2b | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| OR (95% CI) | Wald P | OR (95% CI) | Posterior P | OR (95% CI) | Posterior P | OR (95% CI) | Posterior P | OR (95% CI) | Posterior P | |
| ADRB2 | 1.15 (0.73, 1.81) | 0.555 | 1.22 (0.79, 1.79) | 0.205 | 1.11 (0.76, 1.59) | 0.333 | 1.04 (0.85, 1.27) | 0.368 | 1.03 (0.84, 1.28) | 0.399 |
| CAT | 1.92 (1.18, 3.11) | 0.008 | 1.76 (1.12, 2.64) | 0.006 | 1.72 (1.10, 2.61) | 0.01 | 1.06 (0.88, 1.29) | 0.312 | 1.37 (1.08, 1.77) | 0.005 |
| CC16 | 0.65 (0.42, 1.02) | 0.06 | 0.75 (0.48, 1.09) | 0.064 | 0.72 (0.47, 1.05) | 0.047 | 1.00 (0.82, 1.22) | 0.505 | 0.78 (0.61, 0.98) | 0.017 |
| EPHX1 | 0.96 (0.61, 1.51) | 0.848 | 0.97 (0.63, 1.45) | 0.4 | 1.03 (0.69, 1.48) | 0.478 | 1.07 (0.89, 1.30) | 0.276 | 1.08 (0.86, 1.33) | 0.263 |
| GPX1 | 0.86 (0.55, 1.34) | 0.503 | 0.88 (0.55, 1.31) | 0.251 | 0.89 (0.59, 1.25) | 0.242 | 0.96 (0.79, 1.17) | 0.331 | 0.96 (0.77, 1.17) | 0.349 |
| GSTM1 | 0.95 (0.60, 1.48) | 0.805 | 0.93 (0.60, 1.37) | 0.331 | 0.92 (0.62, 1.40) | 0.304 | 1.07 (0.85, 1.31) | 0.286 | 1.01 (0.79, 1.31) | 0.509 |
| GSTM3 | 0.65 (0.40, 1.06) | 0.086 | 0.67 (0.42, 1.04) | 0.037 | 0.77 (0.48, 1.12) | 0.088 | 0.92 (0.74, 1.10) | 0.172 | 0.92 (0.73, 1.13) | 0.205 |
| GSTP1 | 0.85 (0.54, 1.33) | 0.478 | 0.90 (0.58, 1.33) | 0.273 | 0.93 (0.62, 1.32) | 0.312 | 1.04 (0.85, 1.26) | 0.372 | 1.06 (0.87, 1.29) | 0.288 |
| HO1 | 1.36 (0.86, 2.16) | 0.193 | 1.30 (0.84, 1.96) | 0.138 | 1.46 (0.96, 2.13) | 0.04 | 1.06 (0.88, 1.27) | 0.289 | 1.33 (1.06, 1.67) | 0.007 |
| ICAM-1 | 0.83 (0.48, 1.44) | 0.505 | 0.88 (0.50, 1.44) | 0.26 | 0.80 (0.48, 1.27) | 0.156 | 0.96 (0.78, 1.15) | 0.313 | 0.76 (0.58, 0.96) | 0.01 |
| MMP9 | 1.58 (1.01, 2.46) | 0.045 | 1.47 (0.95, 2.20) | 0.044 | 1.50 (0.99, 2.16) | 0.027 | 1.10 (0.92, 1.34) | 0.167 | 1.27 (1.00, 1.58) | 0.025 |
| NOS3 | 1.13 (0.73, 1.76) | 0.581 | 1.19 (0.76, 1.80) | 0.231 | 1.14 (0.76, 1.64) | 0.284 | 0.98 (0.79, 1.18) | 0.417 | 0.98 (0.80, 1.19) | 0.377 |
| NQO1 | 1.76 (1.11, 2.78) | 0.015 | 1.61 (1.05, 2.44) | 0.016 | 1.56 (1.06, 2.27) | 0.013 | 1.03 (0.86, 1.27) | 0.408 | 1.29 (1.03, 1.62) | 0.016 |
| PPARR | 0.64 (0.37, 1.10) | 0.104 | 0.68 (0.39, 1.13) | 0.06 | 0.74 (0.45, 1.11) | 0.078 | 0.96 (0.79, 1.18) | 0.329 | 0.96 (0.76, 1.19) | 0.342 |
| TGFβ1 | 1.10 (0.70, 1.73) | 0.665 | 1.12 (0.74, 1.66) | 0.324 | 1.08 (0.70, 1.61) | 0.389 | 1.05 (0.86, 1.29) | 0.328 | 0.87 (0.68, 1.12) | 0.123 |
| TNFA | 0.61 (0.37, 1.01) | 0.055 | 0.69 (0.43, 1.04) | 0.036 | 0.63 (0.40, 0.97) | 0.019 | 0.97 (0.79, 1.16) | 0.377 | 0.76 (0.58, 0.95) | 0.01 |
Note: For the standard one-level logistic regression analysis, maximum likelihood estimates of odds ratios (ORs), 95% confidence intervals (CIs), and p value of Wald significant testing were reported in this summary table.
Note: For each of the four hierarchical modeling approaches, posterior estimates of odds ratios (ORs), 95% credible intervals (CIs), and p value were reported in this summary table.
Note: Statisitcal significant findings (two-sided p values less than 5%) were highlighted in red.