Inverse variance-weighted average method

Inverse variance-weighted average method (IVW) is one of the most popular summary-data-based MR methods. While individual-level genetic data are usually considered sensitive and protected, GWAS and quantitative trait loci (QTL) summary statistics are easier to obtain. These summary statistics can then be used to reach a larger sample size and a higher power in a meta-analysis. Here, we assume that the GWAS and QTL summary statistics are generated from different individuals (two-sample setting) and we have effect size estimates, standard error estimates, statistics, and -values for SNPs as IVs. In our setting, the SNPs are at most 1M base pairs apart from the gene body so that they count as cis-QTLs of the gene.

When the genetic variants are uncorrelated, the IVW estimate of the causal effect size of the outcome on the exposure is [6]

where the standard error of the estimate is

Here, is the regression coefficient of the exposure on the SNP in the form of QTL and is the regression coefficient of the outcome on the SNP in the form of GWAS. is the standard error of the regression coefficient of the outcome on the SNP .

The omics biomarker can stand for transcriptomics, proteomics, metabolomics or other omics readouts. This allows us to combine such -values using a method we choose (i.e. Fisher’s combination test, Cauchy combination test, etc.).