Dear all,
I use the Haddock server to run protein protein docking simulations and I really find it very useful. However, I would like to check some results locally. I have downloaded the results of the complete run as a gzipped tar file, as proposed on the server. However, whereas I expected to find all the results (for example all the pdb files corresponding to a given cluster), only the “best” four structures of each ranked cluster are provided if I am not wrong.
In fact, I am surprised by the fact that some clusters are characterized by what appear as good scores (Haddock and z-scores) whereas as the rmsd are significantly high (ranging from 11 to 16 angs). That is why I would like to analyse more deeply the results (of the last two jobs finished on 30/09/25). Is there a way to really access to all the results ?
Many thanks, all the best
JY
I use the Haddock server to run protein protein docking simulations and I really find it very useful. However, I would like to check some results locally. I have downloaded the results of the complete run as a gzipped tar file, as proposed on the server. However, whereas I expected to find all the results (for example all the pdb files corresponding to a given cluster), only the “best” four structures of each ranked cluster are provided if I am not wrong.
In fact, I am surprised by the fact that some clusters are characterized by what appear as good scores (Haddock and z-scores) whereas as the rmsd are significantly high (ranging from 11 to 16 angs). That is why I would like to analyse more deeply the results (of the last two jobs finished on 30/09/25). Is there a way to really access to all the results ?
First of all, the z-score is just a statistical measure telling you where the cluster is with respect to the average score of all clusters.
As for the RMSDs, they are simply measured with respect to the best model generate (the top ranked one).
As as the top ranked model might not have to be per se correct, RMSDs are just indicative.
All the cluster members can be found in various files present in the structures/it1/water directory.
Refer to the manual analysis pages of the haddock online manual:
[
](https://www.bonvinlab.org/software/haddock2.4/analysis/#cluster-based-analysis)
Dear Prof Bonvin,
Well received, many thanks for your fast reply.
For some positions of the protein protein interactions I am investigating (specific arginine residue involved at the interface) the results, for all the clusters obtained through the docking, have the same trends (high rmsd values, ranging from 12 to 22… despite “reasonable” I understand these are pure statistics indications scores). What is also “disturbing” is that if I consider the best four structures of each cluster, they are very close to each other with low (around 1) rmsds. I have rerun the docking to check, similar trends are observed. Would this mean that this position is less relevant at the interface ?
Many thanks for your time,
Best regards
JY
Are you referring to the RMSDs reported by the server?
These are not RMSDs within the clusters, but the average RMSD for that cluster to the top-ranked model from all models generated.
In the structures/it1/water directory you will find among other the following files:
• `file.nam_clustX_rmsd`: contains the RMSD of each structure of cluster X from the best (lowest) HADDOCK score structure of cluster X.
• `file.nam_clustX_rmsd-Emin`: contains the RMSD of each structure of cluster X from the best (lowest) HADDOCK score structure of all calculated structures
Yes, I was referring to the RMSDs reported by the server.
Many thanks for the detailed answer !
