Colocalization

To investigate whether shared genetic variants underlie the causal associations identified in the Mendelian randomization analysis, we conducted a colocalization analysis.

For each metabolite GWAS, we ranked the lead SNPs based on their statistical significance.

To avoid linkage disequilibrium, we iteratively filtered these SNPs, ensuring that each variant was positioned at least 1 Mb apart from any SNP with higher significance.

We then obtained GWAS summary statistics for metabolites and disease traits located within a 1Mb region of each lead SNP.

Prior probabilities were set as follows: P1=1×10−4, P2=1×10−4, and P12=1×10−5, as recommended.

If the posterior probability generated by coloc package with PP.H4>=0.8 was considered strong evidence indicating a common causal variant influencing both metabolites and diseases.

We utilized the coloc package (v5.2.3) in R to perform these analyses.