This talk will advance the notion of gene-environment equivalence as the fundamental principle of Mendelian randomization, from when it was introduced as an approach to strengthen causal inference regarding potentially modifiable factors influencing human development and disease. The biological basis and dimensions of equivalence will be presented, and its relation to the consistency
/treatment–variation irrelevance assumptions in causal inference. The sanity-checking aspects of gene-environment equivalence can be further extended through incorporating recent advances in the molecular biology of gene action.
