Fig. 1From: Causal association of type 2 diabetes with amyotrophic lateral sclerosis: new evidence from Mendelian randomization using GWAS summary statisticsFlowchart displays the selection process for instrumental variables of T2D to investigate the causal effect of T2D on ALS in the Mendelian randomization analysis. A: a set of index SNPs (p < 5.00E−8) were selected to ensure the relevance assumption was satisfied; we further used the F statistic [46, 47] to examine the strength of those index SNPs. B1: following previous studies [48], we attempted to exclude pleiotropic associations by removing index SNPs which were likely associated with ALS with a marginal p value less than 1.00E−5 (the genome-wide suggestive significance level); no index SNPs were removed by this strategy in our analysis; B2: we removed index SNPs which may potentially be in linkage disequilibrium with ALS-associated loci (identified in [49], Additional file 1: Table S7), if the index SNP position was within 1 Mb of an ALS-associated locus. C: similar to B2, based on previous GWAS results of lipid traits [50, 51], we removed index SNPs that were associated with lipids, since dyslipidemia may be linked to both T2D and ALS [45, 52,53,54]Back to article page