[gmx-users] Thin film drifting

[David van der Spoel]

Den 2020-03-29 kl. 19:43, skrev Alex:
> Thanks.
> 
> On Sun, Mar 29, 2020 at 11:59 AM David van der Spoel <spoel at xray.bmc.uu.se>
> wrote:
> 
>> Den 2020-03-29 kl. 15:16, skrev Alex:
>>> Thank Prof. van der Spoel for the response.
>>> No, it isn't. The thin film is solid. There are interaction within the
>> thin
>>> film and with water in the interface.
>>>
>>> Please find a short movie of the unwrapped trajectory of the simulation
>> in
>>> below link (water molecules are hidden); It shows minimization, then
>>> equalization (NVT : tcoupl = v-rescale), then equalization (NpT : tcoupl
>> =
>>> v-rescale and pcoupl = berendsen) and then production (NpT : tcoupl =
>>> v-rescale and pcoupl = Parrinello-Rahman).
>>>
>>> https://drive.google.com/open?id=1jk-Pun1BICNArGJaW0ydIY5TTivECGMf
>>>
>>> The thin film starts drifting significantly in the production along both
>> x
>>> and y directions.
>> Thanks, do you have isotropic pressure scaling?
>>
> Yes, the pcoupltype is isotropic. Please find the mdp file in below link.
> https://drive.google.com/open?id=1iYt6eTcQ4SBi1A9ZFOenL7Q7wc37UvLa
> 
>>
>> Is the shift towards higher x and y values?
>>
> The shift is toward higher y value (+y).
> It also firstly goes toward higher x value (+x ) for a very short time and
> then switches the direction to toward lower x value (-x) and finally shifts
> toward lower x value (-x).
> 
But how is the solid modeled? Are there covalent bonds? It is not 
certain there is a problem, at least I am not convinced. In P scaling 
the coordinates of all particles are scaled by a constant and then 
certain atoms can hop over PBC. If you plot the center of mass including 
periodic boundaries (i.e. without unwrapping) it will be nicely in the 
center of the box with few fluctuations. gmx traj may do this for you.

I and many others have done PMFs of small compounds over membranes of 
different types, but not sure that anyone has noticed problems with 
moving membranes.


> Maybe you can open an issue here:
>> https://gitlab.com/gromacs/gromacs/-/issues
> 
> I will open an issue.
> 
> Actually, in the later steps of the simulations, using umbrella sampling
> and wham, I want to simulate the PMF a single molecule (called A; 26 atoms)
> when it comes from water and diffuses inside the thin film till mid (com)
> of thin film. pull_group1 and pull_group2 would be the Mol_A and thin_film,
> respectively.
> 
> I don't know if this kind of drifting would affect on the PMF, if so, how?
> If the comm-grps works I have the two following options to consider:
> “comm-grps = thin_film Mol_A SOL”
> or
> “comm-grps = Other SOL”
> Where the Other group contains the thin film and molecule A.
> Which one do you recommend, please?
> 
> Thank you,
> Alex
> 
>>
>>>
>>> Best regards,
>>> Alex
>>>
>>> On Sun, Mar 29, 2020 at 2:44 AM David van der Spoel <
>> spoel at xray.bmc.uu.se>
>>> wrote:
>>>
>>>> Den 2020-03-29 kl. 05:24, skrev Alex:
>>>>> Dear all,
>>>>> In a system, I have a thin_film (infinitive in x-y directions) with
>> water
>>>>> on top and bottom of it, PBC = xyz.
>>>>> By the below flags I try to remove the motion of the center of mass of
>>>> the
>>>>> two group separately.
>>>>> comm-grps     = thin_film Water
>>>>> comm-mode   = Linear
>>>>> nstcomm         = 100
>>>>>
>>>>> However the thin film drift specially in x and y directions whereas I
>> was
>>>>> expecting to have no drifting for the thin film, If I understood
>>>> correctly
>>>>> the usage of the comm-grps!
>>>>>
>>>>> Would you please let me know how I can stop drifting of the thin film?
>>>>>
>>>>> Thank you,
>>>>> Alex
>>>>>
>>>> Is that a liquid film? Are there interactions within the film and with
>>>> water? The comm removal will calculate the center of mass taking
>>>> periodic boundaries into account so if your film moves one molecule at a
>>>> time the COM will stay in place. In a realistic system the friction
>>>> between water and film should prevent this, hav eyou tried turning off
>>>> comm?
>>>>
>>>> Historically this has been a fix for the Berendsen thermostat that
>>>> accumulates energy, however with a stochastic thermostat it should not
>>>> be necessary. Not sure about Nose-Hoover though.
>>>> --
>>>> David van der Spoel, Ph.D., Professor of Biology
>>>> Head of Department, Cell & Molecular Biology, Uppsala University.
>>>> Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205.
>>>> http://www.icm.uu.se
>>>> --
>>>> Gromacs Users mailing list
>>>>
>>>> * Please search the archive at
>>>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>>>> posting!
>>>>
>>>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>>>
>>>> * For (un)subscribe requests visit
>>>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>>>> send a mail to gmx-users-request at gromacs.org.
>>>>
>>
>>
>> --
>> David van der Spoel, Ph.D., Professor of Biology
>> Head of Department, Cell & Molecular Biology, Uppsala University.
>> Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205.
>> http://www.icm.uu.se
>> --
>> Gromacs Users mailing list
>>
>> * Please search the archive at
>> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before
>> posting!
>>
>> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>>
>> * For (un)subscribe requests visit
>> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or
>> send a mail to gmx-users-request at gromacs.org.
>>


-- 
David van der Spoel, Ph.D., Professor of Biology
Head of Department, Cell & Molecular Biology, Uppsala University.
Box 596, SE-75124 Uppsala, Sweden. Phone: +46184714205.
http://www.icm.uu.se