Fig. 2

Overview of the two-sample Mendelian randomisation analyses framework. We selected 97 SNPs as instruments for maternal BMI from a genome-wide association studies (GWAS) metanalysis conducted by the Genetic Investigation of ANthropometric Traits (GIANT) consortium [62, 63], including 339,226 males and females. For the selected SNPs, we extracted summary data for the SNP-BMI associations from the GIANT GWAS metanalyses of European ancestry females (N = 171,977) and SNP-outcomes associations from European ancestry females from the MR-PREG collaboration (N range = 92,002 to 446,526). After harmonising SNP-BMI and SNP-outcomes’ summary data, two-sample MR analyses were carried out using the inverse variance weighted (IVW) method, and a series of sensitivity analyses was performed to assess the plausibility of the core Mendelian randomisation assumptions as specified in the figure. For two studies (Generation R and INMA), summary data was only available to us for 32 SNPs reported in an earlier GIANT BMI GWAS [63], of which 12 SNPs overlapped with the 97 selected SNPs and were included in our metanalyses