In other words, Mendelian randomization should be conducted incorporating underlying biology. Clever maths alone will not save you, only bury you deeper.
@AJHGNews @nilanjan10c This shows difficulty of neutrality - the new review by Hu emphasizes robustness to violations, rather than ability to detect a causal effect, heavily favouring a conservative method. In reality, ideal is to rely on biological knowledge.View quoted post on X
