Comparison of Haddock Scores between different Complexes

Is it correct to compare the Haddock scores of docking analyses from different complexes (same protein with different peptide ligands with equal number of residues and a few mutations of a common motif) to determine which “complex is better” and which “complex is worse”?

This is indeed ok, but this is not a ddG (change in binding affinity)

Also it all depends on how you setup you docking (e.g. do you have good information to guide the docking).

An alternative would be, giving a good model or structure of one protein-peptide complex, only use the refinement interface and compare the various mutants.

In addition to that you may run MM/PBSA approach - see dMMPBSA

Thank you very mucj for answering.

This answer leads to another question sorry…

The refinement interface should be used after having generated a model of the complex (e.g. by docking run) in order to optimize a structure already available ? or the the best complex structure generated by refinement interface can be used directly for the comaparison ?

Thank you again

If you have a model or structure of the wild type complex, you could simply make mutations in it and then use the refinement interface. I.e. no docking