Hi,
Actually, for systems beyond simple small molecules my feeling is that
estimating entropy from the QH approximation will be fairly inaccurate
from a quantitative standpoint. Mike Gilson has done a lot of nice
work in this area; you may find this paper useful:
http://pubs.acs.org/doi/abs/10.1021/ct0500904
It discusses in detail using the QH approximation to obtain
configurational entropy from MD simulation data.
-Dan
On Thu, Sep 21, 2017 at 1:16 PM, Mo Rahman
<mo.rahman.postgrad.manchester.ac.uk> wrote:
> Hi,
>
> Thank you very much!
>
> That does clear up a lot of things for me, and I’ve managed to get the calculation to completely successfully. I will be sure to take a look at the tutorials you’ve linked.
>
> Can I ask if you know whether there are any tutorials among those you referred to (or guidance elsewhere) specifically for determining the entropy from MD output using normal mode analysis? I’m not even sure if cpptraj would be the right tool for doing that.
>
> Kind regards,
>
> Mo
>
> On 21 Sep 2017, at 16:54, Daniel Roe <daniel.r.roe.gmail.com<mailto:daniel.r.roe.gmail.com>> wrote:
>
> Hi,
>
> On Thu, Sep 21, 2017 at 11:36 AM, Mo Rahman
> <mo.rahman.postgrad.manchester.ac.uk<mailto:mo.rahman.postgrad.manchester.ac.uk>> wrote:
> First, I believe I need to obtain an average structure, stripped of solvent, and for that I used the the following input:
>
> parm ~/entropy-md/alanine/setup/ala-solv.prmtop #(topology file of the alanine dipeptide in solution)
> trajin ~/entropy-md/alanine/production/ala-solv-prod.mdcrd
> autoimage
> strip :WAT
> average ala-avg.mdcrd crd nobox
>
> Typically you obtain an averaged structure for QH calculations in
> order to RMS best-fit trajectory coordinates to it in order to remove
> overall rotational/translational motion. If this is indeed what you
> want, you also need to have an 'rms' command prior to your 'average'
> command, otherwise your average structure will contain contributions
> from these motions which is probably not what you want.
>
> This completes successfully, however the next step, the actual QH calculation fails to complete. I used the following script:
>
> parm ~/entropy-md/alanine/setup/ala-vac.prmtop #(topology file of the alanine dipeptide alone)
> trajin ala-avg.mdcrd
> createcrd ala-avg.nc
> crdaction ala-avg.nc mwcovar name arcov out arcov.dat
> runanalysis diagmatrix arcov out qh.out thermo outthermo thermo.dat temp 300
>
> This input is definitely not what you want. You're reading your
> average coordinates as the input trajectory, then queuing up an action
> that creates a coordinates (COORDS) data set from that input
> (createcrd), and then trying to use that COORDS data set before its
> even created. I recommend reading the manual so that you understand
> what each of these commands do and become familiar with how Actions
> and Analyses work in cpptraj.
>
> First, you want to generate your covariance matrix from your input
> trajectory (ala-solv-prod.mdcrd), not the average structure. You may
> want to rms-fit to the average structure as I previously mentioned, in
> which case you need to load it as a reference:
>
> parm ~/entropy-md/alanine/setup/ala-solv.prmtop #(topology file of the
> alanine dipeptide in solution)
> trajin ~/entropy-md/alanine/production/ala-solv-prod.mdcrd
> reference ala-avg.mdcrd
> strip :WAT
> rms reference
>
> You then want to create the covariance matrix; this is done via the
> 'matrix' command (which you didn't actually include before, just its
> keywords).
>
> matrix mwcovar name arcov out arcov.dat
>
> You can save the rms-fit coordinates via 'createcrd' if you want to
> later use them for projection in this run, but otherwise it's not
> necessary. You then want to use the 'diagmatrix' command, but there is
> an important distinction to be made. The 'runanalysis <cmd>' command
> means to run the analysis *now*, whereas specifying the analysis
> command by itself will queue it up after any actions, which is what
> you want in this case.
>
> diagmatrix arcov out qh.out thermo outthermo thermo.dat temp 300
>
> In order to use the 'runanalysis' command successfully you would need
> to execute the trajectory processing run first so that the matrix is
> created:
>
> run
> runanalysis diagmatrix ...
>
> Hope this helps.
>
> -Dan
>
> PS - You may find some of these cpptraj tutorials useful:
> http://ambermd.org/tutorials/#analysis
>
>
> The second script seems to fail after createcrd is invoked, and cpptraj reports that the trajectory file contains no frames: the same thing happens when I load in a trajectory with just the solvent removed, but not averaged.
>
> Any help with this would be greatly appreciated.
>
> As an additional request, I would be grateful for any assistance/pointers to guidance for performing normal mode analysis of my output, to obtain entropy using that method as well.
>
> Kind regards,
>
> Mo
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org<mailto:AMBER.ambermd.org>
> http://lists.ambermd.org/mailman/listinfo/amber
>
>
>
> --
> -------------------------
> Daniel R. Roe
> Laboratory of Computational Biology
> National Institutes of Health, NHLBI
> 5635 Fishers Ln, Rm T900
> Rockville MD, 20852
> https://www.lobos.nih.gov/lcb
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org<mailto:AMBER.ambermd.org>
> http://lists.ambermd.org/mailman/listinfo/amber
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
-------------------------
Daniel R. Roe
Laboratory of Computational Biology
National Institutes of Health, NHLBI
5635 Fishers Ln, Rm T900
Rockville MD, 20852
https://www.lobos.nih.gov/lcb
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Sep 21 2017 - 13:00:04 PDT
