Re: [AMBER] Temperature based trajectory in REMD

From: Gargi Borgohai <>
Date: Wed, 2 Oct 2013 14:35:13 +0530

Dear AMBER Users,

Thank you for your suggestions.. I have not found any experimental evidence
for "cold denaturation" of the peptide. But at that temperature (i.e. 267
K), I have carried out one set of conventional MD simulation (using the
same force field viz. amber ff99SB) and the peptide did not show any sort
of denaturation.
Moreover I have carried out cluster analysis of the peptide for temperature
based trajectory (at 267K) in ptraj and I have found that major populations
are occupied by native peptide. Depending on the number of clusters I
assign in the ptraj input script different distributions are obtained and
native is always being the major population. e.g.
Two Cluster Distribution : 84.8%, 15.2%
Three Cluster Distribution : 71.8%, 13.0%, 15.2%
Four Cluster Distribution : 60.5%, 13.0%, 15.2%, 11.3%
Five Cluster Distribution : 55.5%, 13.0%, 5.0%, 15.2%, 11.3% and so on..

Will you please explain how many number of cluster should actually be
assigned? What are the DBI and PSF values written in XXXout.txt file?

Thanking You in advance
Gargi Borgohain
IIT Guwahati

Message: 31
Date: Tue, 24 Sep 2013 11:11:44 -0400
From: Carlos Simmerling <>
Subject: Re: [AMBER] Temperature based trajectory in REMD
To: AMBER Mailing List <>
Content-Type: text/plain; charset=ISO-8859-1

I also suggest doing a cluster analysis before doing this kind of melting
curve. If the "native" is only 10% at low T, it's entirely possible that
your simulation setup (force field, solvent, etc) actually prefer a
different structure. What I mean is - is the 90% that is not native just a
mix of unfolded things, or is it possible that there is a big cluster of a
specific incorrect conformation and the "native" one is a minor conformer?
It's important to know if your experimental structure just isn't stable
enough in the simulation, or if the simulation is predicting the wrong
structure entirely.

On Tue, Sep 24, 2013 at 10:56 AM, Daniel Roe <> wrote:

> Hi,
> On Tue, Sep 24, 2013 at 6:44 AM, Gargi Borgohai <>
> wrote:
> > TEMPLIST=`cat temperatures.dat`
> >
> > for T in $TEMPLIST ; do
> >
> > trajin remd.mdcrd.001 remdtraj remdtrajtemp 267.0
> > trajout remd.Ttraj.267
> This script as written will give only the trajectory at 267 K. It
> should be something like:
> trajin remd.mdcrd.001 remdtraj remdtrajtemp $T
> trajout remd.Ttraj.$T
> FYI you may be interested in trying the 'ensemble' command in cpptraj,
> which will sort and process all temperatures at once. For example, the
> input to cpptraj in your case would be (no 'for' loop required):
> ensemble remd.mdcrd.001
> trajout remd.Ttraj
> You would obtain trajectory files 'remd.Ttraj.X', where X corresponds
> to a position sorted temperature list starting from 0. You can also do
> some simple actions on the entire ensemble at once (calculate
> distances, rmsds etc).
> > While calculating average value of helical fraction
> > (averaged over all frames) for each temperature the value is found to be
> > near about same for each temperature. And when population of "native
> state"
> > was calculated for each temperature, it was found to decrease with
> > increasing temperature (which is expected). But this % of population at
> the
> > lowest temperature (i.e. at 267K) is very small, its about ~10% of the
> > total frames of the whole temperature based trajectory (which is
> > from remd.Ttraj.267). But this is not expected. I may have done some
> > mistake either in finding out temperature based trajectory or in finding
> > the population of "native state" at each temperature. Will you please
> help
> > me in choosing the right pathway to calculate the same? This will be a
> > great help for me.
> Unfortunately there is no easy answer here, and without knowing more
> about what your system actually is I can only speculate. Do you have
> experimental evidence for what the population should be at 267 K? Also
> 267 K is getting pretty cold, and some proteins do undergo a
> phenomenon known as "cold denaturation" (although usually a
> temperatures a little colder). If you started your REMD simulation
> with a conformation other than native it could be that you haven't run
> long enough and your ensemble simply hasn't converged. Do you have at
> least one other independent REMD simulation that gives similar
> results?
> Hope this helps,
> -Dan
> --
> -------------------------
> Daniel R. Roe, PhD
> Department of Medicinal Chemistry
> University of Utah
> 30 South 2000 East, Room 201
> Salt Lake City, UT 84112-5820
> (801) 587-9652
> (801) 585-9119 (Fax)
> _______________________________________________
> AMBER mailing list
AMBER mailing list
Received on Wed Oct 02 2013 - 02:30:04 PDT
Custom Search