Refine definition of 'complex of molecules' to soften the non-covalent binding criterion?

[nataled]

The current definition of 'complex of molecules' contains a clause insisting that the components be bound to each other non-covalently. However, while the vast majority of molecular complexes do indeed rely on non-covalent bonds, that is not always the case. For example, immunoglobulins--which contain inter-chain disulfide bonds--are considered protein-containing complexes in GO.

Should the definition of 'complex of molecules' be slightly reformulated?

The current definition (not counting any modifications currently under consideration) is: "A complex of two or more molecules that are not covalently bound." Possible softening of this wording:

"A complex of two or more molecules that, typically, are not covalently bound." (Perhaps also add a note about notable exceptions such as the aforementioned immunoglobulins). Doing so will allow 'protein-containing complex' to fit comfortably as a subclass of 'complex of molecules'. ('protein-containing complex' currently has this note: "This is not covalently bonded, which conflicts with changes to the parent definition.")