bio-multi-omics-mgwas-integration

An mGWAS workflow involves transforming metabolite concentrations into PLINK phenotypes, running association scans, clumping significant hits, and applying TwoSampleMR for causal inference. This process links genetic variants to metabolite levels.

Colocalization in mGWAS uses the coloc R package to test if a metabolite-associated SNP and a disease GWAS signal share the same causal variant. It assesses whether genetic variants driving metabolite levels also influence disease outcomes.

To prepare metabolomics data for PLINK association testing, convert metabolite concentration matrices into PLINK phenotype files. Apply appropriate phenotype transformation and include covariates before running PLINK 1.9 or 2.0 association scans.

Yes, PLINK 2.0 supports mQTL mapping through association testing and clumping to identify independent loci. After running per-metabolite GWAS, clump genome-wide significant hits to isolate independent mQTLs for downstream colocalization.

For two-sample Mendelian randomization, select instruments from clumped mQTL loci and harmonize summary statistics. TwoSampleMR workflows then estimate causal effects of metabolites on disease outcomes using these harmonized genetic instruments.

Use colocalization to identify shared causal variants between metabolites and disease at a specific locus, and Mendelian randomization to estimate the causal effect of metabolites on disease across the genome. Both address causality but at different scopes.