kb. For the two markers showing the strongest signals in our study, rs8111428 and rs2302382, there are no direct HapMap data available. However, their physical positions indicate that they could be part of the first haplotype block, as confirmed when using our family data in Haploview (data not shown). For the SNP, rs1800437, LD HapMap data was available indicating that this SNP belongs to the second region which was supported by our family data (data not shown). The pairwise r2 values between the SNPs (using our family data) are shown in Table 3. Subsequently, we performed analyses of the transmitted haplotypes (Table 4). One haplotype (estimated frequency = 21%) that included the minor alleles of rs8111428 (G-allele) and rs2302382 (A-allele), was more frequently transmitted in the families (p = 0.003). Testing all haplotype combinations this haplotype had the smallest adjusted p-value of 0.0055 which is corrected for multiple testing. As the haplotype analysis revealed that no haplotype by itself leads to a stronger association signal, we decided to validate the best initial SNPs results.
