From: Raman Preet Singh (ramanpreetsingh_at_hotmail.com)
Date: Wed Aug 12 2020 - 12:44:32 CDT

Dear Josh and Peter,

Thank you very much for the detailed info.

Regards,
Raman

From: Josh Vermaas<mailto:joshua.vermaas_at_gmail.com>
Sent: 31 July 2020 02:56
To: Peter Freddolino<mailto:petefred_at_umich.edu>
Cc: Raman Preet Singh<mailto:ramanpreetsingh_at_hotmail.com>; vmd-l_at_ks.uiuc.edu<mailto:vmd-l_at_ks.uiuc.edu>
Subject: Re: vmd-l: Non-standard solvents in Solvate

Hi Raman,

Adding a few details to Peter's answer:

1.) I've found that the box size equilibrates in about 5ns of NPT. Your mileage may vary.
2.) I find it convenient to have my solvent boxes always be exactly 10000 residues so the resids are 0 through 9999, which fits into the pdb format nicely. The solvent box size doesn't matter. If you need a bigger box, the solvate plugin will replicate the box for you. If you need a smaller box, the solvate plugin will cut the box down to size.
3.) That isn't how the solvate plugin works. It places solvent molecules such that they aren't too close to the solute. If the initial arrangement is too tightly packed, the simulation box will expand under NPT, and if there is too much space, the box will compact. Basically, use NPT for equilibration.
4.) Sure, why not? For mixed liquids, I've made a biphasic system, and then let them mix. It doesn't take very long if the solvents are miscible. For gases, they should explode outward pretty quickly. Indeed, if you just want to simulate gases, the solvate plugin is overkill, since your particles are going to be REALLY far apart anyway, so just initially placing things on a grid wouldn't be a bad approximation.
5.) You just need something that VMD will tag once per residue. If you monkey with your residue definitions in the topology file you are using, you can set the name of one atom in each solvent molecule to be named "SOLV" or something, and then use the key selection "name SOLV". Or just mix single-component solvent boxes through simulation.

-Josh

On Thu, Jul 30, 2020 at 3:08 PM Peter Freddolino <petefred_at_umich.edu<mailto:petefred_at_umich.edu>> wrote:
Dear Raman,
Please see below.

On Thu, Jul 30, 2020 at 4:17 PM Raman Preet Singh <
ramanpreetsingh_at_hotmail.com<mailto:ramanpreetsingh_at_hotmail.com>> wrote:

>
>
> 1. For creating a pdb and psf of the solvent box, I need to
> equilibrate the box of solvent (or solvent mixture). For how long (10 ns or
> more?) and using which ensemble (NPT?) should equilibration be performed?
>
> For most cases, an NPT equilibration would suffice, for long enough for
key thermodynamic parameters (average pressure, temperature, volume,
energy) to stabilize

>
> 1.
> 2. Can the box size of the solvent be different (smaller or larger)
> from the box size of the final system (protein + solvent)? If yes, then how
> much difference is acceptable assuming that the final box size will be
> 50x50x50 angstrom.
>
> Yes, it can be. Solvate will tile your solvent box through space as
needed. I would recommend a 20 or 30 Angstrom side length on the box that
you equilibrate.

>
> 1.
> 2. When a protein is inserted in the equilibrated solvent box, the
> protein will displace the solvent resulting in an increase in pressure