Cohort-wise BMI association of SNPs genotyped in two general Japanese populations.

Effect sizes are indicated as beta per SD unit of phenotype.Two-tailed P values are shown in the table. T2D-risk, fasting plasma glucose (FPG)-increasing and BMI-increasing alleles reported in the previous studies are tested. When we arbitrarily categorized the samples into two age groups (age<60 years and age≥60 years), there was no significant inter-age-group difference in BMI association at CDKAL1 rs4712523; P = 0.076, β = −2.66 for age<60 years; P = 0.006, β = −3.85 for age≥60 years in the combined samples. Part of the samples (414 individuals in the Amagasaki Study panel) were included in the GWA meta-analysis of BMI (ref.35,36), where proxy SNPs (r2>0.78 in HapMap JPT+CHB) were tested for SNP-BMI acid association at CDKAL1. For the purpose of comprehensive evaluation, all the Amagasaki Study samples consecutively-enrolled are included in the present study.