Two-sample summary data MR
Two-sample approaches are a novel extension to 2SLS that greatly increase the scope of MR, because they allow for greater sample sizes and thus greater statistical power. Contrary to a 2SLS approach, in which the gene-exposure and gene-outcome relations are estimated in the same sample, 2-sample approaches derive the estimates from separate samples (e.g., separate GWASs of exposure and outcome) (57, 58).
Two-sample approaches exploit the growing availability of summary data from large consortia of GWASs, such as the GIANT (Genetic Investigation of ANthropometric Traits) consortium, the Global Lipids Genetics Consortium, the International Consortium for Blood Pressure, and the CARDIoGRAM (Coronary ARtery DIsease Genome wide Replication and Meta-analysis) consortium. Table 3 provides a nonexhaustive list of publicly available summary data.
The major assumptions of the approach are that the gene-exposure and gene-outcome associations should be estimated in nonoverlapping samples and should be representative of the same population, practically defined as being of similar age and sex distribution and the same ethnic group.