VMD-L Mailing List

From: Vermaas, Joshua (Joshua.Vermaas_at_nrel.gov)
Date: Fri Aug 23 2019 - 10:39:59 CDT

Hi Neena,

Why not just load the final states sequentially into the same molecule as one trajectory and then use the clustering tools? The clustering tools do not care if the trajectories all came from the same simulation or not. So code like this would probably be appropriate:

set filelist [glob file*.pdb]
mol new
foreach file $filelist {
mol addfile $file
}

-Josh

On 2019-08-23 09:35:00-06:00 owner-vmd-l_at_ks.uiuc.edu wrote:

Hello VMD users,
I tried using the VMD clustering tool and uploaded the cluster.log file from gromacs with a cutoff RMSD for filtering. However, it only allows to extract number of clusters from a single trajectory file at a time.
  What do I want? I have 50 different simulations for the same molecule, but I am not interested in cluster analysis for the trajectory of each simulation. Instead, I want to do cluster analysis on the final 50 structures.
Any insight would be appreciated,
Many thanks,
Neena
________________________________
From: Neena Susan Eappen <neena.susaneappen_at_mail.utoronto.ca>
Sent: Friday, August 16, 2019 4:42 PM
To: vmd-l_at_ks.uiuc.edu <vmd-l_at_ks.uiuc.edu>
Subject: Re: vmd-l: Extract Most Frequently found structure
Thank you Ashar and Joao,
To use the clustering tool in VMD, I just download the plugin from this site? https://github.com/luisico/clustering/blob/master/clustering.tcl
Thank you,
Neena
________________________________
From: Joao Ribeiro <jribeiro_at_ks.uiuc.edu>
Sent: Friday, August 16, 2019 1:11 PM
To: Neena Susan Eappen <neena.susaneappen_at_mail.utoronto.ca>; vmd-l_at_ks.uiuc.edu <vmd-l_at_ks.uiuc.edu>
Subject: Re: vmd-l: Extract Most Frequently found structure

Dear Neena,

As long as these 50 (I am assuming) independent simulations are the same in terms of composition (number and sequence of atoms), you can load then into the same molecule. If for some reason they are different, you need to load them into separate molecules.

Regarding the “number of unique structures/ conformations”, you first need to define what is a unique structure/conformation. RMSD would probably be the first choice, but this calculation can mask a lot of different conformations due to its overall system average. Other measurements are more sensitive to local variations like RMSF per residue, the radius of gyration, contact maps. Depending on the system, you can also define distances and angles between domains as a criterion to distinguish conformations. Once you found the right set of measurements, there are several clustering analysis that you can use.

All of these measurements are possible in VMD, requiring more or less scripting on your end, but first, you need to understand your system.

I hope this helps.

Best,

João

From: <owner-vmd-l_at_ks.uiuc.edu> on behalf of Neena Susan Eappen <neena.susaneappen_at_mail.utoronto.ca>
Date: Thursday, August 15, 2019 at 10:31 PM
To: "vmd-l_at_ks.uiuc.edu" <vmd-l_at_ks.uiuc.edu>
Subject: vmd-l: Extract Most Frequently found structure

Hello VMD users,

I simulated a protein 50 times, is there a way to upload all 50 on VMD and figure out number of unique structures/ conformations? In addition, how to figure out the most frequently found structure?

Many thanks,

Neena