All of the techniques described need appropriate controls. In knockout animals, many investigators will try and 'rescue' the phenotype by adding back the knocked-out protein. Why is this a critical control and what is the expected result, in general?

Knockouts are generated to remove specific proteins. Even under highly controlled conditional knockouts the removal of the protein could have unintended side-effects, for example interfering with another protein. Therefore adding the protein of interest back should fully rescue the phenotype (whatever that is). If adding the protein does not rescue the phenotype then the knockout was not successful or it was not selective.