Table 2 Results of analysis of nonadditive effects
Genotypic modelAdditive modelKORA studyTwinsUK study
SNPMetabolite/ratioA/BAFBeta(BB) (SE)Beta(AB) (SE)P-valueBeta (SE)P-valueLRTAICLRTAICReplicated
rs7552404C12/C10G/A0.30/0.311.150(0.076)0.642(0.046)1.69 × 10−720.600(0.033)1.84 × 10−73aaaa*
rs7601356C9/PC.ae.C30.0C/T0.63/0.65−1.225(0.068)−0.961(0.068)1.24 × 10−70−0.523(0.032)1.57 × 10−58gg/dgg/a*
rs715Gly/GlnC/T0.69/0.70−0.937(0.082)−0.139(0.084)4.28 × 10−69−0.590(0.036)1.00 × 10−61rg/rr,ar*
rs8396C7.DC/C10C/T0.71/0.68−0.760(0.086)−0.360(0.087)5.98 × 10−26−0.388(0.036)4.54 × 10−27aar,ag/r*
rs2046813PC.ae.C42.5/PC.ae.C44.5C/T0.69/0.690.647(0.084)0.360(0.084)6.29 × 10−170.309(0.036)7.11 × 10−18aaa,r,da*
rs273913C5/PC.ae.C34.1T/C0.41/0.350.606(0.071)0.203(0.052)1.60 × 10−160.283(0.034)1.28 × 10−16ag/aa,d,r,oa
rs3798719PC.aa.C42.5/PC.aa.C40.3T/C0.25/0.23−0.983(0.100)−0.418(0.048)5.01 × 10−32−0.453(0.038)6.78 × 10−33aaa,d,ra*
rs12356193C0G/A0.17/0.16−1.145(0.167)−0.483(0.053)2.18 × 10−27−0.507(0.046)2.42 × 10−28aaa,da*
rs603424C16.1/C14A/G0.80/0.820.872(0.121)0.558(0.124)3.70 × 10−180.362(0.041)1.42 × 10−18ag/aa,r,o,da
rs174547PC.aa.C36.3/PC.aa.C36.4C/T0.70/0.65−1.872(0.068)−1.019(0.069)2.29 × 10−208−0.904(0.029)5.98 × 10−209ag/agg/a*
rs2066938C3/C4G/A0.27/0.26−1.942(0.077)−0.649(0.043)1.73 × 10−159−0.832(0.032)7.49 × 10−149gg/agg/a*
rs4902242PC.aa.C28.1/PC.ae.C40.2C/T0.85/0.87−1.019(0.181)−0.356(0.189)2.00 × 10−35−0.619(0.049)3.77 × 10−36a,rg/ar,ar*
rs1077989PC.ae.C32.1/PC.ae.C34.1C/A0.46/0.47−0.888(0.066)−0.529(0.056)6.80 × 10−42−0.450(0.033)1.99 × 10−42ag/aa,dg/a*
rs4814176SM..OH..C24.1/SM..OH..C22.1T/C0.36/0.420.792(0.074)0.422(0.050)2.69 × 10−310.403(0.034)2.00 × 10−32aad,ad*
rs6970485lysoPC.a.C28.0/PC.aa.C26.0C/T0.35/-0.980(0.107)0.877(0.063)1.21 × 10−470.635(0.047)4.66 × 10−41dd
rs1894832Ser/TrpC/T0.51/0.510.491(0.067)0.302(0.061)1.98 × 10−120.245(0.034)4.02 × 10−13aaa,r,da
rs2657879His/GlnG/A0.21/0.190.747(0.128)0.303(0.050)2.89 × 10−140.328(0.042)4.44 × 10−15aad,o,ad*
rs7200543PC.aa.C36.2/PC.aa.C38.3G/A0.31/0.280.531(0.078)0.328(0.049)7.45 × 10−160.287(0.035)1.40 × 10−16a,daa,da*
rs1466448SM.C18.1/SM.C16.1C/A0.22/0.19−0.775(0.120)−0.305(0.049)7.01 × 10−16−0.337(0.041)1.31 × 10−16aaa,da*
rs5746636xLeu/ProT/G0.24/0.270.705(0.103)0.379(0.049)2.98 × 10−200.366(0.039)2.64 × 10−21aaa,r,da*
  • The GWAS results for the genotypic and additive model for 20 identified loci in KORA are reported. Loci that could be replicated in the TwinsUK data are marked with an asterisk in the last column. In the LRT column, all 1-df models are given for which the genotypic model was not significantly better (in descending order of P-value). The genotypic model is indicated in this column if it led to a significantly better fit than all 1-df models. In the AIC column, the best model regarding AIC is named and separated by a slash from the best 1-df model if the best model is genotypic (r, recessive; a, additive; d, dominant; o, overdominant; g, genotypic). The table is divided into two parts: the upper part reports loci that were already found in the previously published GWAS on the same data (Illig et al. 2010); the lower part reports new loci. A, other allele; AF, allele frequency of the effect allele for KORA/for TwinsUK; B, effect allele; Beta(BB) (SE) and Beta(AB) (SE), estimated beta for BB and AB genotypes for genotypic model with standard error; Beta (SE), estimated beta for SNP with standard error.