We can use the genSSI structural identifiability toolbox to do this, and add that as a sub-toolkit withing this toolbox.
One way this could be useful would be in helping determine the simplest model for any given circuit specification. For instance, we already do this in the case of transcription, with multi vs single step mechanisms depending on whether RNA interference is present. Or in the proposed sigma factor mechanism where sigma 70 only gets modeled explicitly when other sigma factors are present. (This is proposed in the issue titled "Removing extraneous reactions #30".)
Another example of conditional models could be to just use single step protein production (instead of splitting it into TX and TL) when all we are modeling is transcription factor mediated regulation.