Table 1

Association results of the 11 loci and MS risk assessed in 21 589 subjects of European descent

This studyOriginal study4Combined
SNPLocation (hg38)Nearest geneNMAFOR (95% CI)p Value*I2 (95% CI)PQORp Value†ORp Value†
rs1534422 (G/A)chr2:12,500,615LOC1005064571144.31.10 (1.06 to 1.15)1.39×10−60 (0 to 56)0.5301.062.49×10−71.074.03×10−12
rs6435203 (G/A)chr2:203,746,472CD281129.51.08 (1.03 to 1.13)7.20×10−432 (0 to 67)0.1421.071.23×10−71.071.35×10−9
rs9846396 (T/C)chr3:141,422,126ZBTB381140.41.03 (0.99 to 1.08)0.05600 (0 to 54)0.5751.077.77×10−81.105.94×10−8
rs4686953 (G/A)‡chr3:188,365,131LPP1145.51.05 (1.01 to 1.10)6.00×10−30 (0 to 52)0.6031.062.26×10−71.063.35×10−8
rs727098 (C/T)chr6:139,590,185DQ5718241024.81.02 (0.97 to 1.07)0.2540 (0 to 53)0.6091.076.49×10−71.062.20×10−6
rs13260060 (A/G)chr8:70,306,125NCOA21110.81.06 (0.99 to 1.13)0.04449 (0 to 48)0.3571.107.40×10−71.092.37×10−7
rs3004212 (T/C)chr10:43,147,362CSGALNACT21126.11.04 (0.99 to 1.09)0.06400 (0 to 0)0.9761.072.99×10−71.063.01×10−7
rs3809006 (G/A)chr11:128,540,941ETS11148.21.07 (1.02 to 1.11)1.26×10−338 (0 to 69)0.09781.063.51×10−71.067.74×10−9
rs806349 (T/C)chr13:50,285,854DLEU11046.71.12 (1.07 to 1.16)9.80×10−854 (6 to 77)0.02131.067.85×10−71.078.14×10−12
rs6072343 (A/G)chr20:41,339,548LPIN3913.71.14 (1.07 to 1.22)1.14×10−538 (0 to 72)0.1121.097.13×10−81.117.16×10−12
rs9808753 (G/A)chr21:33,415,005IFNGR21011.71.12 (1.05 to 1.19)2.62×10−40 (0 to 61)0.4231.091.26×10−71.104.40×10−10
  • Fixed effect meta-analysis results for the SNPs tested across all validation data sets and after combining the results of the original4 and of this study were performed using R. Allele names are displayed as minor/major allele based on control frequencies across all validation data sets. The underlined allele name corresponds to the risk allele. The Bonferroni-corrected threshold for significant association in the validation data sets was set to p=4.5×10−3 to account for 11 tests; upon combining all data sets from the original4 and validation study, the threshold to define significance was set to p=5×10−8, that is, corresponding to genome-wide significance. Underlined p values indicate experiment-wide significance for the validation data sets or genome-wide significance for the meta-analysis of all available data. The locations are annotated based on the human genome build 38 (hg38) and the nearest gene has been determined according to RefGene as annotated on the UCSC (University of California, Southern California) Genome Browser. Note that the nearest gene does not necessarily represent the functional element underlying the genetic association.

  • *One-sided with respect to the effect direction of the original study,4

  • †Two-sided.

  • ‡This SNP was listed with archive ID rs66756607 in the original publication.4

  • I2, estimate of percentage of between-study heterogeneity that is beyond chance; MAF, minor allele frequency in controls across all validation datasets (in %); MS, multiple sclerosis; N, number of validation data sets after quality control; PQ, p value derived from Q statistic to assess between-study heterogeneity.