(~8% in the US). It is interesting to note that the estimate of bxy at the TCF7L2 locus strongly deviated from those at the other loci (Fig. 3), suggesting that the TCF7L2 SNP has pleotropic effects on BMI and T2D. The TCF7L2 SNP was detected as an outlier by the HEIDI-outlier method and removed from the GSMR analysis. In addition, the risk effect of BMI on asthma is in line with the result from a recent MR study (using a weighted genetic allele score as the instrument) that higher BMI increases the risk of childhood asthma36. Moreover, we identified a protective effect of BMI against osteoporosis (OR = 0.68), consistent with the observed associations in previous studies37,38. The estimated risk effect of BMI on T2D in the community data (OR = 3.29) was similar to that in the case–control data (OR = 3.12, Fig. 2b and Supplementary Data 1). We also observed a strong risk effect of BMI on coronary artery disease (CAD) in the case–control data (OR = 1.70), in line with the risk effect of BMI on CVD (OR = 1.30) in the community data.
