Error in HADDOCK3 Clustering (RMSD-based [clustrmsd]) during Protein-Protein Docking

Hi,
I’m currently using HADDOCK3 on an HPC for my protein-protein docking, and I’ve opted for the RMSD-based clustering method to cluster my models [clustrmsd]. However, I encountered an error during the docking process. The rigidbody, flexref, emref, and mdref stages run smoothly, but I encountered an issue specifically with clustrmsd.: [2024-02-21 18:02:49,331 init INFO] Running [seletop] module
[2024-02-21 18:02:49,512 init INFO] Module [seletop] finished.
[2024-02-21 18:02:49,512 init INFO] [seletop] took 0 seconds
[2024-02-21 18:02:49,603 init INFO] Running [clustrmsd] module
[2024-02-21 18:02:49,605 libutil ERROR] list index out of range
I keep the default parameters, and this is how I have them set :
[clustrmsd]
criterion = ‘maxclust’
linkage = ‘average’
threshold = 4

hi,

you should add a matrix calculation module before clustrmsd, like rmsdmatrix or ilrmsdmatrix. something like this in your workflow:

[seletop]

[ilrmsdmatrix]

[clustrmsd]

My advise is to use rmsdmatrix if you have a good knowledge of the interface, as you can specify it using the resdic parameter (an example here haddock3/examples/docking-protein-glycan/docking-protein-glycan-test.cfg at main · haddocking/haddock3 · GitHub ). if your knowledge of the interface is not great then ilrmsdmatrix is a better option (see haddock3/examples/docking-protein-glycan/docking-protein-glycan-ilrmsd-test.cfg at main · haddocking/haddock3 · GitHub).

Hope it helps!

Hi Marco,
thanks for getting back to me quickly. I’m currently working on ab-initio protein-protein docking, so I don’t have any information about the interface or interaction surface. I tried checking the ilrmsdmatrix module by running “haddock3-cfg -m ilrmsdmatrix,” but I couldn’t find the module. Is it available on HADDOCK3? Can caprieval provide similar metrics without needing a reference structure? As I mentioned, I’m running in ab-initio mode.

that means that you have to update your HADDOCK codebase (see haddock3/docs/INSTALL.md at main · haddocking/haddock3 · GitHub). In the case of ab-initio docking the ilrmsdmatrix is indeed the best choice.

as the code is still evolving fast check also the parameter names once you update the codebase (we changed the names of some of them lately)

Hi @marco.giulini ,
can I still use haddock3 for my docking ever if the code hasn’t finished yet ? will the code still has a major modification ? does the available code sufficient to use haddock3 ?
Thanks to you for all the responses that you have given me.

The code is still evolving, but close to the first production release.

Simply make sure to update your installation in case you run into problems.