Hi William,
This sounds very strange to me. My first instinct would be that the force field had not been converted over to NAMD properly, or something is wrong with NAMD settings - SCEE / SCNB scaling etc. However, the fact that during the MD you see things appear to fix themselves suggests something else.
My guess would be the following:
You ran minimization and heating / equilibration with constant volume and then you switched to constant pressure for the production run. Am I correct? If yes that would explain it perfectly - you likely have vacuum bubbles forming during the constant volume phase. If yes see if you might be able to specify the box size a little better and try switching to constant pressure as soon as you possibly can - ideally as soon as the temperature hits 100K or so during the heating.
All the best
Ross
> On Jan 28, 2015, at 8:13 AM, William Hester <william_hester.me.com> wrote:
>
> Hi,
>
> I have built a large system using the charmm membrane builder which consists of a large ion channel protein embedded in a POPC lipid bilayer. With the ions and water molecules the system is about 150 000 atoms in total. I am running the simulations on NAMD using AMBER forcefield. The membrane was successfully converted into AMBER format. However, when running the minisation of the system, the membrane separates and the two layers move away from each other leaving a gap in the middle of the system. During the heating and equilibration, this problem also persists. However, I ran 1ns of molecular dynamics to see how the membrane behaves and it actually readjusts itself back to the correct configuration.
>
> Has anyone seen or heard of this before? I’ve doubled checked the configuration files and boundary conditions and that is not the problem. I ran three minimisations, the first with fixed membrane and protein, the second with fixed protein, and the third with nothing fixed. I also tried to run the heating without minimising the membrane as the charmm GUI is supposed to produce an equilibrated structure. However, the membrane layer separation still persisted in this case.
>
> Any insights would be greatly appreciated.
>
> Regards
>
> Will
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Jan 28 2015 - 19:30:02 PST
