[gmx-users] [OT] Software package for discontinuous molecular dynamics? (DMD)

[David van der Spoel]

On 2/2/10 10:53 PM, ms wrote:
> David van der Spoel ha scritto:
>> On 2/2/10 4:43 PM, ms wrote:
>>> David van der Spoel ha scritto:
>>>
>>>> On 2/1/10 4:32 PM, ms wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> Sorry for the offtopic but Google/literature quick search is not
>>>>> helping
>>>>> and I'd like to have some more informed opinion.
>>>>>
>>>>> To my understanding, GROMACS isn't capable of discontinuous molecular
>>>>> dynamics. Is there any more-or-less standard software package for that?
>>>>>
>>>>> thank you!
>>>>> m.
>>>>>
>>>> Can you be more specific please?
>>>>
>>>>
>>> Uhm, in what meaning?
>>> I was pondering about using DMD for use with a coarse-grained model and
>>> I wondered if there is a standard package for that. Don't know how to be
>>> more specific, what is unclear in my question?
>>>
>>> thanks,
>>> M.
>>>
>> What is discrete MD?
>
> Uh, sorry. Probably references know better than me:
> http://pubs.acs.org/doi/full/10.1021/ja0539140 is an example of it
> applied to proteins. (I am looking for reviews on the subject but can't
> find recent ones) It is a coarse-grained technique for use with
> discontinuous potentials, which roughly speaking uses "collision
> detection" instead of fine-grained potential calculation:
>
> "In a discontinuous molecular dynamics (DMD) simulation, particles
> collide when they arrive at a discontinuity in the potential, that is,
> the hard-sphere diameter or the square-well width. Between collisions,
> particles move with linear trajectories, making DMD simulations much
> faster than traditional molecular dynamics simulations with continuous
> potentials which require a small integration time step. The
> postcollision velocities of particles in DMD are found by solving the
> collision dynamics equations analytically."
>
> (from http://pubs.acs.org/doi/full/10.1021/la049267s )
>
> I don't know much about it, but I was curious and I wanted to know if
> there was a package for it (or if Gromacs was capable of it).
>
> thanks!
> m.
I see, in hindsight I had heard of it on a recent conference. Out of the 
box this will not work in Gromacs. How much work it would be to 
implement I don't know either, and since you lose the advantage Gromacs 
has on fast evaluation of the forces, you may be better off with another 
package.

-- 
David van der Spoel, Ph.D., Professor of Biology
Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205. Fax: +4618511755.
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se