From: Malay Ranjan Biswal (malay_at_jncasr.ac.in)
Date: Thu Jul 13 2017 - 11:49:15 CDT
Thank you Joshua and Jerome for your kind reply. I recenter the membrane every restart of the simulation to prevent the membrane going towards the upper or lower end of the box.
Thanking you
Malay Ranjan Biswal
Theoretical Science Unit
Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR)
Jakkur, Bangalore-560 064, India
----- Original Message -----
From: "Joshua Vermaas" <Joshua.Vermaas_at_nrel.gov>
To: namd-l_at_ks.uiuc.edu, "Malay Ranjan Biswal" <malay_at_jncasr.ac.in>
Sent: Monday, July 10, 2017 11:19:11 PM
Subject: Re: namd-l: Maintaining PBC on Collective Variables
Hi Malay,
There are two options I've tried, one is just to add z-restraints to
selected lipid atoms. This prevents the membrane from moving too far so
it doesn't split, and has the benefit of not needing to be repeated. The
other option is to occasionally recenter the restart file so that the
coordinates remain centered on the membrane before the PBC box is split.
Ideally, what would happen is that the colvars routine can be told to
specify a specific atom or group as the center of the PBC cell, rather
than using NAMD's box center, but I don't think this is in the cards yet.
-Josh
----- Original Message -----
From: "Jérôme Hénin" <jerome.henin_at_ibpc.fr>
To: "Namd Mailing List" <namd-l_at_ks.uiuc.edu>, "Malay Ranjan Biswal" <malay_at_jncasr.ac.in>
Sent: Monday, July 10, 2017 6:50:55 PM
Subject: Re: namd-l: Maintaining PBC on Collective Variables
You can't let your bilayer get split across the PBCs. Two options:
- disable wrapping
- prevent vertical diffusion by restraining the center of mass of the bilayer along the z axis with a distanceZ coordinate. If you do that, use few atoms, eg. one per lipid. You don't even need to restrain all lipids.
I would also try the zeroMomentum option, which should be sufficient to prevent overall diffusion in the first place.
Jerome
On 07/10/2017 07:25 AM, Jérôme Hénin wrote:
> Dear all namd users,
>
> We are doing a lipid bilayer simulation, where we use metadynamics to study insertion of molecules into the membrane. During the simulation, we observe that the membrane translates leaving an apparently split membrane in the minimum image, i.e, lower half of membrane on the top of the box, upper half at the bottom, and water in the middle of the frame. We are using the colvar module to define the z-distance (“distanceZ”) and radius of gyration (“gyration”) as the collective variables for the calculation. However, the PBC is not maintained for these CVs. But by default the “forceNoPBC = off” is active as seen in the log-file. Both “distanceZ” and “gyration” fail to have the PBC correction. For example, due to the membrane splitting the COM of the lipid bilayer lies in water, rather than lying inside the bilayer.
>
> Can someone please suggest how PBC can be maintained on the CVs and this problem be corrected?
>
> Thanking you
>
> Malay Ranjan Biswal
>
> Theoretical Science Unit
> Jawaharlal Nehru Centre for Advanced Scientific Research (JNCASR)
> Jakkur, Bangalore-560 064, India
> https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.jncasr.ac.in&data=02%7C01%7CJoshua.Vermaas%40nrel.gov%7C8423b5974bf8446dad9008d4c7955715%7Ca0f29d7e28cd4f5484427885aee7c080%7C1%7C0%7C636352891598672322&sdata=FI1oeDRFnCBYnxqyRh1rJ2Hq1Y3nnR3PZLrlGQ0tLMY%3D&reserved=0
>
>
This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:21:27 CST