Hello,
All are good questions.
On May 11, 2012, at 3:17 AM, Dilraj LAMA <dilrajl.bii.a-star.edu.sg> wrote:
> Dear Amber users,
> I am currently planing to use temperature based REMD
> for my work. For this I am learning the application following the REMD
> tutorial mentioned in the Amber tutorial
> http://ambermd.org/tutorials/advanced/tutorial7/.
>
> I have few queries and will be glad if you can kindly give me your input.
>
> 1) The first is regarding the chirality restraints. I am planing to
> use the temperature range between 300-400K for my study. I have done a
> search of the mailing list and came across threads
> (http://archive.ambermd.org/200901/0203.html) where it was mentioned
> that for this range of temperature, it is not critical to have the
> chirality restraint. So I am planing to not use it for my study. Is
> this a fair judgment?
In my experience not having chirality restraints over this temperature range is fine.
>
> 2)In the REMD tutorial of Amber, there is an equilibration run
> performed for all the replicas before actually doing REMD. Now in the
> step after the equilibration, the authors have given an example of the
> REMD.mdin (http://ambermd.org/tutorials/advanced/tutorial7/). I am
> just appending the parameters from the input file that is bothering me
> below.
>
> irest=0, ntx=1,
> temp0=XXXXX, ig=RANDOM_NUMBER,
>
> REMD uses the restart file from the equilibration run. So shouldn't it
> be that irest=1 and ntx=5 so that the velocity information is also
> read from the restart file?
When using the restart files from previous MD, use irest=1 and ntx=5 if and only if you wish to use the old velocities (I usually do).
>
> And why are the authors setting the random number and initial
> temperature again for the REMD run of the tutorial when the
> temperatures are already set in the previous equilibration run?
Each replica needs its own temperature, so that value will change for each mdin file that controls a different replica.
>
> 3) I also have a query on the "ig" parameter. In the classical MD,
> "ig=-1" is now routinely set in the input parameter file. Shouldn't it
> be the case also for REMD simulations?
Yes, but ig=-1 was not around when the tutorial was written. Setting ig=-1 is the appropriate thing to do.
HTH,
Jason
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri May 11 2012 - 06:00:06 PDT
