Triangulation

The collaboration uses genetic and non-genetic methods including MR, colocalisation, multivariable regression, paternal negative controls, and within-family analyses. Consistent findings across triangulated approaches strengthen causal interpretation because unrelated biases are unlikely to produce identical result patterns. When triangulated findings disagree, prespecified expectations about bias can guide sensitivity analyses and complementary methods. MR-PREG uses triangulation by integrating evidence from methods, data sources, and study designs with distinct sources of bias.