Comparing ligands with the same protein in docking

I read on this forum that the comparison of binding affinities of ligands different complexes is not reliable. Does this mean that if you have different ligands but the same protein, you can´t compare the performance of the ligands? This is surprising to me since I thought this was the entire purpose of virtual screening, i.e. comparing the binding affinity of different ligands for the same biological target. Could someone please explain this further to me?

Hi Leo

For different ligands with the same protein you can indeed compare results.
But the score does not tell you the exact binding affinity. And I would be careful in saying one ligand is much better than another.

Okay, so you can´t quantify the difference but you can say that the results indicate a difference?

Yes - that would be a safe assumption

We did that actually for a drug repurposing screen against the Sars-Cov2 protease - check bonvinlab.org/covid

Here I think it is more relevant to look for interesting compounds in the top25, top55… than state that number #1 is much better than the others.