Skip to main content
. Author manuscript; available in PMC: 2013 Jan 1.
Published in final edited form as: Ann Rheum Dis. 2011 Oct 6;71(1):136–142. doi: 10.1136/annrheumdis-2011-200085

Table 1.

Summary of the two-gene interactions between BANK1 and BLK in three independent sets of cases and controls

Gene SNP1 Genotype SNP1 Gene SNP2 Genotype SNP2 Population set OR OR L95 OR U95 P Se Pe Freq. Cases Freq. Controls a b c d N
BANK1 BLK
rs10516483 CC rs1478895 CC North Europe 100k 2.38 1.69 3.36 4.83E-07 8.9 0.0001 35% 18% 88 167 92 416 763
Set 1 (USA) - - - - - - - - - - -
Set 2 (Europe) 1.42 1.17 1.72 0.0003 −0.3 0.8256 25% 19% 261 773 285 1200 2519
Meta-analysis b 1.60 1.36 1.89 7.61E-12 nc 349 940 377 1616 3282
rs10516487 GG rs1478895 CC North Europe 100k 1.82 1.35 2.45 8.27E-05 3.7 0.0063 48% 33% 131 144 171 342 788
Set 1 (USA) 1.33 1.06 1.66 0.0124 2.7 0.0170 38% 31% 233 388 241 533 1395
Set 2 (Europe) 1.33 1.13 1.57 0.0005 −1.0 0.9578 42% 35% 434 605 527 980 2546
Meta-analysis a 1.40 1.24 1.58 1.97E-10 nc 798 1137 939 1855 4729
rs10516483 CC rs2736340 TT+TC North Europe 100k 2.32 1.44 3.76 4.68E-04 4.6 0.0111 24% 12% 56 178 31 229 494
Set 1 (USA) - - - - - - - - -
Set 2 (Europe) 1.70 1.36 2.13 3.65E-06 6.3 0.0024 18% 12% 184 824 169 1287 2464
Meta-analysis b 1.80 1.47 2.21 1.18E-11 nc 19% 12% 240 1002 200 1516 2958
rs10516487 GG rs2736340 TT+TC North Europe 100k 1.82 1.22 2.72 0.0031 2.6 0.0373 32% 21% 80 170 54 209 513
Set 1 (USA) 1.57 1.22 2.02 4.36E-04 −1.6 0.9850 27% 19% 168 453 148 626 1395
Set 2 (Europe) 1.62 1.34 1.95 3.55E-07 6.7 0.0013 29% 20% 293 720 297 1181 2491
Meta-analysis a 1.63 1.41 1.87 1.75E-15 nc 29% 20% 541 1343 499 2016 4399
BANK1 GG rs10516487 Meta-analysis a 1.37 1.22 1.54 4.33E-08 56% 48% 4791
CC rs10516483 Meta-analysis b 1.52 1.31 1.76 1.66E-12 36% 27% 3334
BLK TT+TC rs2736340 Meta-analysis a 1.33 1.21 1.46 4.31E-14 30% 24% 8862

The epistasis P value (Pe) measures only the statistical significance of the epistatic effect of the interaction. For every 2 SNP-genotype combination tested, a contingency table under the null hypothesis of independence between both SNPs (no epistasis) is derived and an expected Pearson S score is calculated (S0). The epistatic score is then defined as Se = S − S0. This score is the difference of two dependent scores, each one following asymptotically a one-degree of freedom chi-square distribution. Pe are empirically determined by permutations (100,000 case/control label shuffling). nc: Due to the nature of the Pe statistic, no permutation p-value can be calculated for meta-analysis.

North Europe 100k: discovery set, genotype data extracted from a 100.000 SNPs genome-wide scan in Swedish. The SNP rs2736340 was genotyped a posteriori. Replication Set 1 (USA): European-American set. Replication Set 2 (Europe): A combined set of German, Italian, European-Argentine and Spanish cases and controls.

a = cases with genotype SNP1 AND genotype SNP2, b = cases with other genotype, c = controls with genotype SNP1 AND genotype SNP2, d = controls with other genotype. N: number of cases and controls with non-missing genotypes for both SNP 1 and SNP 2. Frequencies of the 2 SNP-genotype combinations in cases and controls are shown.