[gmx-users] problem with trjconv -pbc cluster

[jagannath mondal]

Hi Chris,  Thanks a lot for your responses. I will surely try them. But, before I go to trjconv -pbc cluster, I had one more problem to sort out. This has to do with quantifying the cluster-size distribution from the trajectory. As I mentioned, if I visualize the trajectory in VMD, I see that starting from a random dispersed state of 50 molecules of surfactant, it aggregates gradually and at a long time, two distinct miceller aggregates are formed.But, If I try to quantify the cluster-size distribution using g_clustsize utilty, it gives me bizarre results: the maxclust.xvg for the final time frame shows that I have 1 single cluster containing all the 50 molecules. But, Vmd shows 2 distinct aggregate.
Here is the command line I used:   g_clustsize -f traj.xtc -s topol -mol -cut 0.50 I used -mol option so that I get the cluster size distribution in terms of the molecules.Here, executing the commands gives me maxclust.xvg which has maximum size of clusters( in terms of molecule number) as a function of time frame. It shows at the end, maxclustersize  is 50 and it has all the particles in it( as obtained from maxclust.ndx)Besides, at any other time, the maximum size of particles predicted by g_clustsize does not match with what VMD shows. I also  tried post-processing the trajectory using trjconv with -pbc whole or with -pbc nojump . But, g_clustsize always returns same result:maxcluster contains all the particles.Then I tried playing with -cut option : I started with default value of 0.35, here the result is worse: it shows maxclust = 1 i.e all the particles are monomeric . But any value beyond 0.35 gives me maxclust = 50.   I tried to
 cross-check whether visualization is misrepresented or not: for this I tried to calculate all the distances among the com of the surfectants using g_dist tool and tried to manually cluster the surfectants and I found that here I can easily come up with two separate cluster based on the distances. So, I think that the problem lies in g_clustsize not in visualization, as -pbc whole or -pbc nojump does not change the result.So, do you have any suggestions ? Any help in pointing out where I am doing wrong will be helpful.Jagannath

--- On Wed, 7/7/10, chris.neale at utoronto.ca <chris.neale at utoronto.ca> wrote:

From: chris.neale at utoronto.ca <chris.neale at utoronto.ca>
Subject: [gmx-users] problem with trjconv -pbc cluster
To: gmx-users at gromacs.org
Date: Wednesday, 7 July, 2010, 9:27 PM

PS:

you realize that trjconv -pbc cluster is never going to work unless you actually have a cluster, right? So you can't start this procedure near your initial dispersed state. You need to start at some time T where you actually do have a cluster. Then you can try -dump T, -dump (T+X), -dump (T+X*2), etc. Then you can follow the ides expressed below like running backwards with -pbc nojup to get the initial state.

Chris.

--original message --

Dear Jagannath:

There is, as far as I know, no way to fix this. Here's what you should do.

1. Let the space between your timesteps be X ps.
2. Use trjconv -pbc cluster -dump X -o out.gro
2b. if that doesn't work, try trjconv -pbc cluster -dump (X*2) -o out.gro
2c. if that doesn't work, try trjconv -pbc cluster -dump (X*3) -o out.gro
...
(Note use real numbers for args to -dump).

Once you have a frame that works, you can run part 2 and 3 of that
wiki page by making a tpr based on the gro that worked -- and I think
that you'll need to run your trjconv -pbc mol starting from the frame
where clustering worked (e.g. -b X). Note that this second requirement
means that you may need to reverse the frames from your .xtc file so
that you can run in reverse time from a complete micelle through to
disassembly with -pbc nojump -- a for loop around trjconv will work
but inefficiently, you might be able to use -demux smartly here (I'm
not sure).

But then since you have 2 micelles you will need a starting box in
which they are both whole.... getting trickier. You should be aware
that the trjconv -pbc cluster is not the only way to do part 1 from
that wiki page. You could use trjconv -trans X Y Z and then make a new
.tpr and then run trjconv -pbc nojump and it would work if only you
knew which X Y Z to use (although trial and error may help you here
eventually).

Sorry this is confusing, but this is a difficult thing to do and you
should expect to struggle with it for some time, so another questino
to ask yourself is "do I really need to do this?" If its just for
making a movie then you can use pbctools in vmd for that.


Chris.

-- original message --

Hi,  I had a system of surfectants which are started with an initial
configuration where all of them are well dispersed. Visual study of
trajectory shows they start aggregating and finally form two discrete
micelles. To quantify this micelle clusterization, I tried to use the
suggestions present in the gromacs documentations:
http://www.gromacs.org/Documentation/How-tos/Micelle_Clustering
i.e    use trjconv -pbc cluster to obtain a single frame that has all
of the lipids in the unit cell. This must be the first frame of your
trajectory. A similar frame from some previous timepoint will not
work.use grompp to make a new tpr file based on the frame that was
output from the step above.use trjconv -pbc nojump to produce the
desired trajectory using the newly produced tpr file.trjconv -f a.xtc
-o a_cluster.gro -e 0.001 -pbc clustergrompp -f a.mdp -c a_cluster.gro
-o a_cluster.tprtrjconv -f a.xtc -o a_cluster.xtc -s a_cluster.tpr
-pbc nojump
But, The first step i.e. trjconv -pbc cluster does not work. Looks
like it is going through an infinite loop and is not stopping for
convergence.
I am using gromacs 4.0.7
Any help will be appreciated.
Jagannath




--
gmx-users mailing list    gmx-users at gromacs.org
http://lists.gromacs.org/mailman/listinfo/gmx-users
Please search the archive at http://www.gromacs.org/search before posting!
Please don't post (un)subscribe requests to the list. Use the
www interface or send it to gmx-users-request at gromacs.org.
Can't post? Read http://www.gromacs.org/mailing_lists/users.php


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20100707/04d3e7ce/attachment.html>