It might seem surprising that GIPR has not been detected in any of the currently published genome-wide association studies on obesity (BMI) [e.g. [26,44-48]]. There are at least two explanations: First, genome-wide SNP chips do not cover well the region of GIPR. For example, the Affymetrix Genome-Wide Human SNP 6.0 Array (with more than 906,600 SNPs) in the region ± 50 Kb of GIPR only comprises 3 SNPs and none of these SNPs is within GIPR. The Illumina 580 K array on the other hand comprises 19 SNPs in the region ± 50 Kb of GIPR, but only three of them are within the gene. One idea would be to use imputation analyses to solve this problem. Imputations, however, heavily rely on just a few HapMap individuals and the assumption that linkage disequilibrium between markers is the same for these individuals and the individuals actually genotyped. Moreover, for markers like rs8111428 and rs2302382 there is no HapMap information available making imputations impossible. Second, current meta-analyses of genome-wide association studies did focus on BMI in the general population; it might well be that GIPR variants have a major impact in extremes of the phenotype only.
