AMBER Archive (2006)

Subject: Re: AMBER: simulate a self-assemble process in vacuum

From: Eric Hu (list.eric_at_gmail.com)
Date: Thu Jul 20 2006 - 16:43:35 CDT

On 7/13/06, Gustavo Seabra <gustavo.seabra_at_gmail.com> wrote:
> Hi Eric,
>
> In princliple, I don't see anything wrong with that. However, in
> practice, things are a bit different...
>
> For one, what is the timescale of this process? Most "long" MD
> simulations go up to the nanoseconds timescale (someone please correct
> me if I'm wrong). If the self-assembly process happens in the
> miliseconds timescale, your calculation may take forever (and a bit
> more!) to see eomething happening.

So far we don't know the timescale of the assembly. I assume that
normal md in room temperature cannot get to the final state easily
which is why I need to speed up the sampling process without
introducing too much bias.

> Second, does that self-assembly process involve bonde
> breaking/forming? If so, you will need some sort of QM/MM calculation.
> Amber 9 can do that, but now the timescale you can reach with the
> calculation is even shorter. (Or you'll need "more forevers" to see
> anything happening :-) )

No, there is no bond breaking/forming and only formation of hydrogen
bonds between subunits which competes with the subunit stacking.

> But there is still hope... You may be able to get around this by using
> some steered molecular dynamics, where you give the atoms a little
> push in the right direction, and that can get around the timescale
> problem. However, now you have to be careful because you'll be
> introducing a bias in your system, and you may end up making happen
> things that wouldn't happen otherwise, so a closer understanding of
> what you're trying to simulate becomes even more important.

This is something I'm afraid of since I really have no clue how the
molecules assemble themselves.

> Another possibility is by using replica exchage MD, which will help
> your system overcome barriers. Amber can do that as well. But in this
> case, depending on your system's size, you may need a large number of
> processors and maybe even some special techniques.
>

I do not know too much about this but will have to give it a try.

> In summary, simulating what you want is not impossible, just involves
> a bit more work than just "putting-all-in-one-pot-and-wait".
>
> I hope that helps. (Also, I hope someone here will correct me if I
> said something wrong.)
>

Thanks a lot, Gustavo. I guess I will go ahead to make a pool of
molecules at random positions and see what happens.

Eric

> Best wishes,
>
> Gustavo.
>
> On 7/12/06, Eric Hu <list.eric_at_gmail.com> wrote:
> > Hi, I am trying to simulate a self-assemble process of an organic
> > molecule. My logic is to prepare a series of molecules that have
> > random orientations in high concentration in a limited space. I wonder
> > if AMBER can help achieve this goal. The reason I ask is that in my
> > mind this is a very similar way as creating periodic boundary
> > conditions for your own solvent which some people might already
> > encounter. Thanks!
> >
> > Eric
> > -----------------------------------------------------------------------
> > The AMBER Mail Reflector
> > To post, send mail to amber_at_scripps.edu
> > To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
> -----------------------------------------------------------------------
> The AMBER Mail Reflector
> To post, send mail to amber_at_scripps.edu
> To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber_at_scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo_at_scripps.edu