VMD-L Mailing List

From: Brian Radak (brian.radak_at_gmail.com)
Date: Fri May 25 2018 - 12:40:32 CDT

Deleting waters like that seems like more trouble than it's worth. If you
don't delete exactly the same number of molecules in each window, you would
almost certainly have problems in computing a correct PMF and maybe even
hysteresis issues with ABF. The slow and careful steered MD method seems
fairly recommendable and has worked for us in the past. Unfortunately you
only have your own chemical intuition available when evaluating whether or
not you are in fact being slow and careful.

Cheers,
BKR

On Fri, May 25, 2018 at 11:30 AM, McGuire, Kelly <mcg05004_at_byui.edu> wrote:

> Thank you for the response! I always worried about biasing with deleting
> water molecules at each new position when I was shown umbrella sampling for
> the first time. I would like to avoid that if possible.
>
>
> *Kelly L. McGuire*
>
> *PhD Scholar*
>
> *Department of Physiology and Developmental Biology*
>
> *Brigham Young University*
>
> *LSB 3050*
>
> *Provo, UT 84602*
>
>
> ------------------------------
> *From:* Giacomo Fiorin <giacomo.fiorin_at_gmail.com>
> *Sent:* Friday, May 25, 2018 7:09:32 AM
> *To:* McGuire, Kelly
> *Cc:* VMD Mailing LIst
> *Subject:* Re: vmd-l: ABF Calculations Question 1
>
> Hi Kelly, ABF, metadynamics, umbrella sampling or any of the other methods
> provided in Colvars will only affect directly the atoms that you select,
> and never touch the others. If you want water molecules to move, you will
> have to define variables that involve them explicitly. In particular,
> atoms and molecules are not deleted: that's something you can only do
> during system preparation.
>
> As for how each enhanced sampling method goes about sampling the
> collective variables space, you should take a look at the literature.
> These papers are usually referenced in each tutorial.
>
> Technical information about how to use them in NAMD is here:
> http://colvars.github.io/colvars-refman-namd/colvars-refman-namd.html
> and here:
> https://github.com/Colvars/examples
>
> Lastly, knowing the system that you are looking at, my advice is to look
> at what kind of transformations need to occur for this calculation to
> work. You have three constriction regions, each of which needs to open to
> let the inhibitor molecule pass. Even for smaller movements, some
> wetting/dewetting is necessary, which is not always fast. Explicitly
> adding or deleting water molecules may give the appearance that you are
> sampling wetting/dewetting transitions, but if these don't happen during
> simulation the artificial initial conditions bias the results.
>
> Giacomo
>
> On Thu, May 24, 2018 at 1:10 PM, McGuire, Kelly <mcg05004_at_byui.edu> wrote:
>
> I am preparing to do some ABF calculations with the molecules I sent
> through FFTK by working through the tutorial. I am placing these molecules
> in the influenza A M2 protein to get their free energy profiles in order to
> understand their entry/exit and interactions inside the protein. I will
> have some questions as I work my way through the tutorial. My first one
> is more of a general question on how ABF works. I've done some umbrella
> sampling before, and the way we have done this in the past is create
> multiple input files with the molecule/drug at different positions in the
> protein, delete the waters at the drugs new position, and submit all of
> those jobs to the supercomputer, which as you can imagine is very
> tedious/inefficient. I am wondering though how the ABF set up in NAMD
> handles moving the drug and deleting waters around the drug at the new
> position? For example, say I have the drug starting at one end of the
> protein and we call that position on the reaction coordinate -15 Angstroms,
> and the drug is to be moved to the other end of the protein, call it 15
> angstroms (so, 30 angstroms for the total reaction coordinate), how
> does the part of the colvars script, (width 0.1, lowerboundary -15,
> upperboundary 15), move the drug to the next spot and delete waters around
> it? Is this some kind of a loop that NAMD recognizes to move the drug?
> Thanks!
>
>
> *Kelly L. McGuire*
>
> *PhD Scholar*
>
> *Department of Physiology and Developmental Biology*
>
> *Brigham Young University*
>
> *LSB 3050*
>
> *Provo, UT 84602*
>
>
>
>
>
> --
> Giacomo Fiorin
> Associate Professor of Research, Temple University, Philadelphia, PA
> Contractor, National Institutes of Health, Bethesda, MD
> http://goo.gl/Q3TBQU
> https://github.com/giacomofiorin
>