This variant was associated with being protective vs. obesity, with a p-value in the initial association for obesity in adults of 2.6 * 10^-6 under an additive model. In a replicate follow-up this had a p-value of 2.4 times 10-3. The odds ratio in the replicate adults was 0.71, and 0.75 in obese children. They associate this with a per-allele average BMI difference of -0.087. In a 5’ 6” individual this corresponds to 0.54 less pounds on average per allele.
To estimate attributable decreased risk, we use data from Swiss adults in supplementary table 3, as the controls represent a random selection rather than lean individuals. Obese adult cases have an allele frequency of 0.308, while random adults had an allele frequency of 0.411. Assuming an incidence of obesity of 10%, this corresponds to an attributable decreased risk of 2.3%.
Review article of obesity GWAS studies, including Meyre et al.