You are here

About align hundreds of pdb structures produced by rosseta membrane abinitio

6 posts / 0 new
Last post
About align hundreds of pdb structures produced by rosseta membrane abinitio


I have hundreds of pdb structures produced by rosetta membrane abinitio from one sequence.
And I cluster them to get three clusters.
Now I want to align the structures within each clusters, so that I can get RMSD for each cluster.

I used to use MultiProt to align 3D structures, but seems it has a problem dealing with hundreds of structures.

Dose anyone know a better tool (or feature within rosetta, which I'm not aware) to dealing with alignment for many structures?

Thanks for your help!

Post Situation: 
Wed, 2013-05-22 12:37

If you've clustered them with Rosetta, you should have gotten an RMSD to the cluster center with the output. This should be the aligned and superposed rmsd value.

Failing that, you can use the score application ( to get the rmsd to a given structure. Just pass the reference structure to the -in:file:native command line option.

If you're looking for an all-against-all rmsd for all the cluster members, I'm not aware of anything in Rosetta that would do it (although it would be simple enough to put together in PyRosetta). One possibility is to look into the Calibur clustering program ( & ), which many in the Rosetta community are using for clustering, though I haven't used it myself, so I can't say what information is available.

Wed, 2013-05-22 12:50


I guessed I didn't say it correctly. I used "Calibur" to cluster my pdb structures, produced by rosetta.
Does anyone know how to show the RMSD for each clusters in Calibur? Because I'm pretty sure it calculated RMSD first in Calibur in order to get these largest clusters.

Thanks for your help!

Tue, 2013-05-28 12:20

I emailed the authors of Calibur, who are currently working on incorporating the program/library into the Rosetta source. You should hear from them soon.

Mon, 2013-06-03 10:03


We are the authors of Calibur. We hope that you will find the
following program useful.

The program will accept two parameters:
1. a target_pdb_path: that is, the path of the centroid PDB file
2. a file_list_path: the path to a file that lists out the paths to the
PDB files that are to be aligned with the centroid PDB.

The output is a list of decoy-vs-rmsd. That is, the decoy against
its rmsd to the centroid. You can obtain the total rmsd by summing
up all the rmsd, perhaps using Excel. For each decoy, the program
will also output a file which contains its coordinates when it is
optimally aligned with the centroid.

Please let us know if you would prefer the program to behave
differently. Thanks.

Tue, 2013-06-11 18:46

Could you please share the sourcecode or linux and macos binaries for this application? Thank you.

Tue, 2015-09-22 08:35