MR Dictionary

The process by which potentially disruptive influences on normal development from genetic (and environmental) exposures are damped or buffered by compensatory developmental processes. This results in animals (including humans) being relatively "plastic" during developmental periods. For example, hypothetically fetuses with genetic variants that result in on average higher glucose levels might develop other compensatory systems that lower glucose levels or such that higher glucose levels do not adversely affect them. 

It is not directly possible to measure or determine. If canalization has occurred, then genetic instrumental variables (IVs) may still relate to the exposure by the same magnitude (as in the absence of any canalization) but the effect on a potential outcome of that exposure could be dampened. Biological evidence for canalization having a marked impact on genetic associations is lacking. This will not impact MR results when assessing intrauterine effects (i.e., via maternal genetic IVs for maternal pregnancy exposures) on offspring outcomes, as the mother’s genotype will not have been influenced by canalization in the offspring (though this has additional potential problems. See MR for testing developmental origins). 

References

Other terms in 'Sources of bias and limitations in MR':