Interpreting a large, highly converged cluster in AIR-guided docking

HADDOCK
1

I am docking two protein fragments that I suspect interact via short linear motifs. I used default docking parameters and two models taken from reviewed UniProtKB entries. For this run I defined AIRs on a 12-residue stretch on each protein that I believe represents the interacting motifs.

507×344 6.74 KB

Runs with smaller and larger AIR definitions converge to the same surface, but only this AIR set produces such a large cluster.

Is this result capable of arising from incorrectly placed AIRs, or is this generally taken as evidence for a well-defined docking solution in HADDOCK?

2

It is nice in general if the top cluster is highly populated, but does not have to be case.

When you increase of decrease your number of restraints, considering 50% is randomly deleted for each model generate, you can expect differences in the number and size of the clusters generated.

But if all top ranked clusters are rather consistent in their binding mode, it is a good indication that your runs do converge, irrespective of the exact definition of AIRs.