Previously identified regions and their validations in our study.

a: NCBI Build 36. b: p-value of the most significant haplotype from paper. c: SNAP search criteria: R2>0.8 SNP dataset: T (1000 Genome Pilot 1), H3 (Hapmap3), Distance limit: 10 kb, Population: CEU. d: p-value of the most significant haplotype block with SNPs associated with previously identified haplotypes. P-values are calculated by Haploview with the Solid Spine analysis method (D′>0.6). *: p-values are calculated by creating custom blocks using Haploview software. **: p-values are calculated using the Haploview solid spine method (D′>0.6). The most significant haplotype block of region B is different in CATIE, GAIN and nonGAIN.