I am trying to dock molecules with the same parent core but different substitution groups into the same PDB structure using ROISE.
I tried to read through some published papers on ResettaLigand but still was not able to find out the answer for:
Since we are only allowed to include on ligand in each run, can we submit a few different runs and compare the interface score directly?
I am not sure about the algorism on whether the residue within a certain distances are consider or only interacting ones are considered. I guess in the latter case, we should not compare directly as different residues were used for scoring.
If someone could give a professional answer for this question, I would really appreciate your help.
You have the general sense of it. Typically when we do this sort of virtual screening approach, we run a number of different RosettaLigand runs, and then compare the interface energies.
Note, though, that there are some limitations with comparing interface scores. These scores haven't really been normalized for different sized ligands, or ligans which vary in which functional groups they contain, so they might not be completely accurate in how they rank different ligands in binding.
That said, it's likely the best you'll do in the current Rosetta framework. And experience has shown that it at least does a decent job at ranking compounds. (It does about as well as other such docking programs.) You just have to keep in mind that you will likely get odd outliers (compounds which dock anomolously good/poorly for their experimental binding energies), and for compounds that are close in binding energy you probably won't get a clear signal, but the general trends tend to be there.