Cluster decoys after running PlexPepDock

[Anonymous]

Hello everone

After running the Plexpepdock, I have 40000 decoys. I try to use the cluter program of Rosetta 3.4 . But, I do not understand the meaning of cluster. Because , after I tried to run with 200 decoys and get the out:file with 200 file c.0.XXX.pdb format. I do not understand how cluster work.

Flags:
-in:file:silent decoys.silent
-database /home/tuongvy/SW/rosetta3.4/rosetta_database
-in:file:fullatom
-score:weights
-score:patch
-cluster:sort_groups_by_energy

In my case, I want to device 40000 decoys base on score, and group decoys have near score on a group. Can I use cluster program?

Thank you so much!

[Anonymous]

There are many threads which might be helpful to understand why clustering is important and how it is done in Rosetta. In FlexPepDock, we cluster a small fraction of the decoys. Normally top scoring 1% ( so top 400 decoys based on reweighted score is this case ) decoy is chosen for this purpose. After clustering each cluster is ranked based on energy and top 10 ranking clusters are chosen as representative clusters.
Regarding the output c.i.j means that pdb is the jth member of the ith cluster.

[Author]

Posts