TIP4P

[n-salvi]

This is somewhat related to an old closed issue #590

What is the most convenient way of building a system with a 4P water model, for example for the sake of running simulations in acemd? I tried plenty of options in the past including starting with both charmm and amber force field, with no luck. Eventually I ended up running my simulations in gromacs, but with the new ACEMD3 is would be interesting to try again.

Is there any way of perhaps using Parmed to migrate input files from gromacs to acemd?

[j3mdamas]

Hi @n-salvi,

I recently made a PR that solved that issue: #594

That PR included changes that override Amber default paths. Check the new amber.build argument teleapimports: https://github.com/Acellera/htmd/blob/master/htmd/builder/amber.py#L220

If it does not work, can you share with us a system so we can try it out?

João

[n-salvi]

Hi @j3mdamas,

Thanks, the link you provided seems to be what I am looking for on the amber file.
Is there anything similar for the CHARMM builder?

Many thanks,

Nicola

[j3mdamas]

No, for charmm.build we do not have it.

[stefdoerr]

Hi @n-salvi did you manage to build the TIP4P system? If so we could close this issue.

[n-salvi]

I gave it a try but did not work straight out of the box for me. Sorry I didn't have time to try to understand the issue in details but I intend to do so when I come back from holidays. Of course, if you have a working example to share it would be a tremendous help.

[n-salvi]

Il ven 2 ago 2019, 09:30 Stefan Doerr <notifications@github.com> ha scritto:

…

Hi @n-salvi <https://github.com/n-salvi> did you manage to build the TIP4P system? If so we could close this issue. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#888?email_source=notifications&email_token=AC63FY34UPCSFX2CY62YQPLQCPO75A5CNFSM4HY7BSA2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3M446Q#issuecomment-517590650>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AC63FY43CITCUDBIJFDLHZTQCPO75ANCNFSM4HY7BSAQ> .

[stefdoerr]

I gave it a try with CHARMM and managed to build and run the system but it doesn't really work yet due to psfgen not supporting LONEPAIR statements yet.

Once they are supported, this should work:

First of all we don't include the TIP4P parameters in HTMD. You need to download them from the CHARMM parameter zip and then modify the stream file a bit to only contain a single parameter section.

toppar_water_ions_tip4p.str.tar.gz

Then you need to remove the TIP3P from the default parameters of building by explicitly passing the other default parameter files as below.

Lastly you need to disable the angle and dihedral regeneration of psfgen which I added in this PR because it creates angles between the lone pair "atom" and the other real water atoms and the simulation will complain that there are no parameters for that angle.

from htmd.ui import *
from htmd.protocols.equilibration_v2 import Equilibration

mol = Molecule('3ptb')
mol.filter('protein')
smol = solvate(mol)
smol.resname[smol.resname == 'TIP3'] = 'TIP4'
bmol = charmm.build(smol, ionize=True, outdir='/tmp/test/', stream=['./toppar_water_ions_tip4p.str'], param=['par/par_all36_prot.prm', 'par/par_all36_lipid.prm', 'par/par_all36_cgenff.prm'], regenerate=None)

eq = Equilibration(_version=3)
eq.write('/tmp/test/', '/tmp/test_run/')

This builds and runs fine at the moment but is probably very wrong since the LONEPAIR statements are ignored. Once there is a new psfgen release we will update this issue.

[stefdoerr]

there seems to be lots of active development currently on psfgen:
https://www.ks.uiuc.edu/Research/vmd/vmd-new/devel.html

I think with version 2.0 of psfgen in VMD 1.9.4 we should be able to do it.

[n-salvi]

Thank you @stefdoerr, this is very helpful. I think we need to wait for psfgen 2.0 then...