Re: Using harmonic walls on water molecules

From: Jérôme Hénin (jerome.henin_at_ibpc.fr)
Date: Tue Feb 05 2019 - 03:27:22 CST

I second Josh's answer. What you are trying to do is easy to describe in
terms of many individual variables (coordinates of water molecules), and
quite a bit harder to put in terms of a small number of collective
variables. Here the collective variable you'd want is the number of water
molecules occupying a certain region of space, and that's a non-trivial
thing to do, especially in a way that scales. TclBC is designed just for
that, for scalability.

Jerome

On Tue, 5 Feb 2019 at 02:53, Vermaas, Joshua <Joshua.Vermaas_at_nrel.gov>
wrote:

> Hi Moises,
>
> The pre-colvars way of doing this should still work. TclBC is the one that
> springs to mind first. Below implements something close to what you
> intended.
>
> tclBC on
> tclBCScript {
> proc calcforces {step unique K} {
> if { $step % 20 == 0 } {
> cleardrops
> }
> while { [nextatom] } {
> #Drop things outside of a serial list
> if { [getid] < 179210 || [getid] > 224830 } {
> dropatom
> continue
> }
> set pos [getcoord]
> set z [lindex $pos 2]
> if { $z < 46 && $z > -36 } {
> dropatom
> continue
> }
> if { $z > 47 } {
> addenergy [expr {$K * ($z-47) * ($z-47)}]
> addforce [vecscale [list 0 0 -1] [expr
> 2.*$K*($z-47)]]
> } elseif { $z < -38 } {
> addenergy [expr {$K * ($z+38) * ($z+38)}]
> addforce [vecscale [list 0 0 -1] [expr
> 2.*$K*($z+38)]]
> }
> }
> }
> }
> tclBCArgs {5}
> The downside here is that the forces applied are not relative to the rest
> of your system, but rather in the laboratory/simulation coordinates
> reference frame, so you'd need to do something to fix the protein
> orientation within the simulation cell.
>
> -Josh
>
>
> On 2019-02-04 18:20:16-07:00 owner-namd-l_at_ks.uiuc.edu wrote:
>
> Hi All,
> I am trying to confine a large amount of water molecules within an annulus
> of a protein which is in a vacuum.
> The water molecules escape along the z-directions.
> I have tried to accomplish this by defining harmonic walls using the
> collective variables modules.
> My first attempt involved defining all waters as a single collective
> variable:
> colvarsTrajFrequency 100
> colvar {
> name water
> lowerboundary -39.0
> upperboundary 48.0
> lowerWallConstant 5.0
> lowerWall -38.0
> upperWall 47.0
> upperWallConstant 5.0
> distanceZ {
> main { atomNumbersRange 179209-224829 }
> ref { atomNumbersRange 224830-316341 }
> axis { (0.0 , 0.0 , 1.0) }
> }
> }
> The waters still escaped.
> My second attempt involved selecting each of the ~15,000 waters as an
> individual collective variable; however, this will take a long time to
> initialize the simulation.
> What is smarter way to go about this? Thanks!
> --
> Moises Romero
> PhD Student
> Andricioaei Group
> UC Irvine
>
>

This archive was generated by hypermail 2.1.6 : Tue Dec 10 2019 - 23:20:27 CST