[gmx-users] molecular dynamics problems

vlad.chindea at pharmatech.ro vlad.chindea at pharmatech.ro
Tue Jul 20 14:17:48 CEST 2004

Dear Xavier

First of all, if I am not wrong I know that decreasing thermostat coupling
constant allow faster equilibration of the system. Default value in Sybyl
is also 100fs but when I saw that local temp. are not what they should
be,I decrease the value, in order to heat more the system.

Actually I have to run in NPT, but when I saw that I cannot achieve
equilibrium, and the volume keeps increasing (I used again 20fs coupling
constant with scaled velocities, maybe this is the culprit?) ruining my
density, I switched to NVT, to have a looksee.

Local temp. are mesured by Sybyl by selecting molecules of interest,
(probably by the same algorithm as total temp.) so I choose the whole
solvent molecules and while reported temp. for the system was 300K for
solvent was only 200K.

When I measured the temp. of the solute molecule (only 41 atoms from 4850
of the whole system)it was even lower then solvent temp., which led me to
belive that I have a 'hot solvent, cold solute situation' From your
experience what is the normal variation of local temp.(compared to total
temp.) in a well behaved simulation?

Velocities were assigned randomly at the begining of the simulation.

Bug in the software was more like a rethorical question, I know that most
probably I am doing something wrong but i cannot figure out what.

What would be the normal course that you would employ for a simulation of
this kind (solute in water, equilibrated at 300K in NPT) ?

Many thanks and kind regards
Vlad chindea

 vlad.chindea at pharmatech.ro wrote:
>>Dear GROMACS users
>>Although I do not use this program for the moment I have run into a
>>molecular dynamics problem upon which I need your expert advice. I
>>performed a series of molecular dynamics simulations with Sybyl 6.91 in
>>NTV ensamble.
> Why NVT ?? I guess you checked the density of your system ! If you
> actually
> need to run in NVT perhaps an NVE simulation would be nice to equilibrate
> the volume first.
>>I was simulating a small molecule inside a box with 1600
>>molecules of water, after 20 picosecond of simulation the temperature of
>>the system was reported as 300K as intended but when I measured the local
>>temperature of the whole solvent system, the average was around 200K.
> What do you mean by the "local temperature of the solvent system" ? 20
> ps may be
> too short equilibration ! Did you assign velocities at the start of the
> simulation ?
>>Because the solvent accounted for more than 99% of the mass of the
>> system,
>>I expected a value of solvent temperature of nearly 300K.
> You mean 200K ? If the program gives you 300 K for the whole system that
> means
> that the solvent is probably at 300K. Perhaps check your calculation of
> the solvent alone.
>>Another 10ps of
>>simulation did not change the situation. Thermostat coupling constant was
>>20 femtoseconds, and implicit thermostat in Sybyl is, I think, Berendsen.
> 20 fs constant coupling to the thermostat is way to small !! May be
> better to try 0.1 ps
>>Could you be so kind and inform me: is there a bug in temperature
>>calculation or my expectations were wrong considering the situation ?
> This is unlikely, I guess the programm has been used many time and the
> algorithm to get the temperature is not that fancy to make mistakes on it.
> XAvier
> --
> ----------------------------------------------
>    Xavier Periole - Ph.D.
>    Dept. of Biophysical Chemistry / MD Group
>    Univ. of Groningen
>    Nijenborgh 4
>    9747 AG Groningen
>    The Netherlands
>    Tel: +31-503634329
>    Fax: +31-503634800
>    email: x.periole at chem.rug.nl
>    web-page: http://md.chem.rug.nl/~periole
> ----------------------------------------------
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.

More information about the gromacs.org_gmx-users mailing list