Table 2. Analysis of MYO9B, PARD3 and MAGI2 SNPs in the Dutch and German cohorts and joint analysis.
|
Initial study
|
Follow-up study
|
Joint analysis
|
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RAF patients (n = 387) | RAF controls (n > 800) | P-initial* | RAF patients (n = 235) | RAF controls (n = 250) | P follow-up* | P-joint | OR | 95% CI | P-adjusted | |||
| rs7259292 | MYO9B | T/C# | 0.046 | 0.026 | 0.0053 | 0.047 | 0.020 | 0.0200 | 0.0003 | 1.94 | 1.35-2.78 | 0.0031 |
| rs2305767 | MYO9B | A/G | 0.620 | 0.557 | 0.0021 | 0.590 | 0.607 | 0.5880 | 0.0211 | 0.85 | 0.74-0.98 | 0.1709$ |
| rs1545620 | MYO9B | C/A | 0.448 | 0.364 | 2.3x10-5 | 0.385 | 0.359 | 0.4083 | 5.9x10-5 | 1.33 | 1.16-1.53 | 0.0006 |
| rs1457092 | MYO9B | A/C | 0.401 | 0.337 | 0.0011 | 0.344 | 0.345 | 0.9807 | 0.0062 | 1.22 | 1.06-1.40 | 0.0557 |
| rs2305764 | MYO9B | A/G | 0.433 | 0.381 | 0.0093 | 0.383 | 0.401 | 0.5516 | 0.0614 | 1.14 | 0.99-1.31 | 0.4193 |
| rs10763976 | PARD3 | A/G | 0.483 | 0.431 | 0.0195 | 0.564 | 0.536 | 0.3857 | 0.0157 | 1.19 | 1.03-1.38 | 0.1320 |
| rs4379776 | PARD3 | A/G | 0.371 | 0.312 | 0.0046 | 0.353 | 0.341 | 0.6995 | 0.0109 | 1.22 | 1.05-1.41 | 0.0929 |
| rs6962966 | MAGI2 | G/A | 0.493 | 0.458 | 0.1114 | 0.481 | 0.567 | 0.0077 | 0.8450 | 0.99 | 0.85-1.14 | 1.0$ |
| rs9640699 | MAGI2 | A/C | 0.390 | 0.384 | 0.7708 | 0.391 | 0.413 | 0.4956 | 0.8844 | 0.99 | 0.85-1.15 | 1.0 |
| rs1496770 | MAGI2 | A/G | 0.407 | 0.400 | 0.7317 | 0.391 | 0.409 | 0.5538 | 0.9575 | 1.00 | 0.86-1.15 | 1.0 |
OR, odds ratio; 95% CI, 95% confidence interval; RAF, risk allele frequency.
# Risk variant/second allele
* Two-tailed P-values were calculated by chi-squared test for independence of allele counts
$Heterogeneity between the cohorts (Breslow-Day test).
The risk variant was the associated allele in the Dutch cohort; the same variant frequencies were reported for the German cohort. A combined analysis of the Dutch and German results was performed using Cochran-Mantel-Haenszel analysis with 50,000 random permutations. This generated two P-values (P-joint and P-adjusted), an OR and 95% CI. P-joint shows an individual SNP’s significance in the combined cohort. P-adjusted was obtained after correcting for multiple testing. SNPs rs2305767 and rs6962966 showed modest evidence for heterogeneity between the Dutch and German cohort when a Breslow-Day test was performed on this data.