I have a question about whether my approach towards analysing the haddock results is appropriate. For my work, I am comparing three protein ligands of different animal origins (A, B, C) with two receptors of different animal origins (C, D). If I were to compare only the clusters with the highest HADDOCK score, I would not have meaningful conclusions. Of note, when comparing the different clusters in the same run, two clusters give similar ligand-receptor complex shapes with the ligand shifting slightly in position. The clusters have very different HADDOCK scores and cluster sizes. I noted this observation happening for some runs.
Since there is a reference complex of this ligand-receptor interaction on PDB, I plan to use PyMol to superimpose possible results of the same run to this reference complex. The cluster that fits the best to this reference complex, in terms of RMSD value, I would use this cluster for further analysis (comparing the different ligand and receptor interactions). Is this approach wrong?
Do let me know if I need to clarify myself any further. Thank you!