[jcranmer]

You have to at least build the compound to do the assay to actually find out that it's useful.

[wiggles_md]

What you describe is the current practice. What I am trying to suggest is that it is reasonable to be able to use a Markush claim for obvious related compounds.