Re: Inquiry about lipid membrane simulation

From: Wonpil Im (wonpil_at_ku.edu)
Date: Wed Jun 18 2014 - 14:53:31 CDT

What Sunhwan said is true.

Wonpil

On Jun 18, 2014, at 2:46 PM, Sunhwan Jo <sunhwan_at_uchicago.edu> wrote:

>> - You seem to be using CHARMM-GUI (which you should have mentioned when first asking your question). CHARMM-GUI generates NAMD input files to allow its users to run production with NAMD. Is that what you're doing, or are you modifying the protocol? And in the latter case, why?
>>
>> I started with CHARMM-GUI, and made some changes mostly in setting up the systems, as I built the force field of a new molecule that is not available in the CHARMM force fields. I did not make much changes in the equilibration and production run files of CHARMM or NAMD from CHARMM-GUI, except that I added some constraints on dihedral angles of the new molecule.
>>
>> Comparing the CHARMM and NAMD equilibration and production input files from CHARMM-GUI, I found that (1) NAMD input did not contain the lines to keep water out from the lipid hydrophobic core in the Equilibration part, while CHARMM input did; (2) CHARMM input uses a small potential of MMFP GEO Cylinder ... for the solute, while NAMD did not. So I wonder if the shrinking result came from these differences.
>
> I don't think MMFP GEO has significant impact on the system size during the equilibration. I think CHARMM-GUI developers are throughly checking the NAMD / CHARMM inputs and were able to reproduce system size, order parameter, etc. You may want to check with developers. I've cc them in this e-mail also.
>
> Thanks,
> Sunhwan
>
> On Jun 18, 2014, at 2:25 PM, Mo Chen <mochen.mmm_at_gmail.com> wrote:
>
>> Hi Kenno,
>>
>> I am sorry I should have emailed to the NAMD list earlier. My answers to your question is as below.
>>
>>
>> On Wed, Jun 18, 2014 at 12:02 PM, Kenno Vanommeslaeghe <kvanomme_at_rx.umaryland.edu> wrote:
>> As a matter of policy, I do not answer this kind of questions per personal e-mail (nor do I debug people's input files in their stead). Additionally, I don't even know the answers to most of your questions; other people on this list do. The only thing I can offer is more questions:
>> - does the membrane shrinkage also happen when you start your NAMD simulation from a CHARMM equilibrated structure?
>> - what lipid(s) does your membrane consist of, and at what temperature are you simulating it?
>>
>> The two sets of simulation were started from the same system. That is, I used CHARMM to do equilibration and production, and parallelly I used NAMD to do equilibration and production. I did not use the equilibrated structure of CHARMM for NAMD production run, but I will definitely try that.
>>
>> - You seem to be using CHARMM-GUI (which you should have mentioned when first asking your question). CHARMM-GUI generates NAMD input files to allow its users to run production with NAMD. Is that what you're doing, or are you modifying the protocol? And in the latter case, why?
>>
>> I started with CHARMM-GUI, and made some changes mostly in setting up the systems, as I built the force field of a new molecule that is not available in the CHARMM force fields. I did not make much changes in the equilibration and production run files of CHARMM or NAMD from CHARMM-GUI, except that I added some constraints on dihedral angles of the new molecule.
>>
>> Comparing the CHARMM and NAMD equilibration and production input files from CHARMM-GUI, I found that (1) NAMD input did not contain the lines to keep water out from the lipid hydrophobic core in the Equilibration part, while CHARMM input did; (2) CHARMM input uses a small potential of MMFP GEO Cylinder ... for the solute, while NAMD did not. So I wonder if the shrinking result came from these differences.
>>
>> Thank you very much.
>>
>> Best,
>> Mo
>>
>>
>>
>> Please answer these on the NAMD list; the information will help other members in trying to find the cause of your problem.
>>
>>
>>
>> On 06/18/2014 02:41 PM, Mo Chen wrote:
>> Hi Kenno,
>>
>> Thank you very much for your prompt response.
>>
>> I was only doing regular MD simulations, nothing special. I did the
>> simulation both in CHARMM and NAMD for a short term (10ns), and found that
>> CHARMM kept the unit cell dimension while NAMD did not, and the NAMD
>> trajectory showed that it converged to a much smaller area/lipid than the
>> standard value proposed in the CHARMM36 force field. Therefore, I figured
>> that maybe my NAMD configuration files were not correct (attached). As I
>> compared the inputfiles of CHARMM and NAMD, I found that NAMD input file
>> lacked the MMFP GEO Plane for water during equilibration stage and MMFP
>> GEO cylindrical for partially inserted protein segment during production
>> run to avoid protein drafting.
>>
>> May I ask if any molecule should be constrained during production runs?
>> Thank you very much.
>>
>> Best,
>> Mo
>>
>>
>> On Wed, Jun 18, 2014 at 8:20 AM, Kenno Vanommeslaeghe
>> <kvanomme_at_rx.umaryland.edu <mailto:kvanomme_at_rx.umaryland.edu>> wrote:
>>
>> On 06/17/2014 09:57 PM, Mo Chen wrote:
>>
>> 1) In CHARMM, the MMFP GEO PLANE ... is used to keep the
>> non-crystal water
>> molecules away from entering the hydrophobic core of lipid bilayer
>>
>>
>> I don't know what kind of study you're doing, but generally spoken,
>> that shouldn't be necessary, and is likely harmful. If, OTOH, the
>> function of the MMFP restraint is to keep the membrane aligned with
>> the box, then something can be said in favor of it (though I still
>> would do it differently).
>>
>>
>>
>>
>

This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:22:30 CST