Re: [AMBER] Restart

From: Henry Green <>
Date: Mon, 6 Sep 2021 19:15:34 +0530

Update on the Restart Problem : it has been observed that using NTX = 5
instead of 7 (for irest > 0) - intended to instruct the engine to read
coordinates and velocity data from the last .ncrst file seems to return the
temperature normally without any drop ( i.e. Temperature continues to
fluctuate about 300 K without dropping). In this regard I have come to know
from my Senior that the use of restraints did not change the temperature
drop; however using ntx = 5 has solved the problem on their system as well.
Apart from that, I was also told that this problem did not occur with the
PMEMD module of AMBER 18 (i.e. ntx = 7 did not cause the temperature drop)
and it seems the problem is being encountered only with ntx = 7 on the
SANDER module of AMBER 18. I am attaching some of the files that I was
shown by my Senior concerning this. Guidance regarding this is greatly

Thanking you,

On Mon, 6 Sept 2021 at 11:05, Henry Green <> wrote:

> Respected Sir/Madam,
> My name is Diganta Dasgupta and I am a student in JNU, India. I have
> recently started to learn MD using AMBER. I am writing this mail to inform
> you of an error which I have encountered while doing so.
> I have started using AMBER 18 and was trying out the Tutorials
> provided at the AMBER website. In the Section 5 of Tutorial no. (5.2)
> titled "Simulating a DNA polyA-polyT Decamer" provided by Ross Walker Sir,
> which involves simulating the Decamer system in explicit solvent, it was
> mentioned that one should heat the system (to 300 K), after minimization
> (2-step), over a run of 20 ps in the NVT ensemble (using SHAKE on the
> Hydrogens and a weak restraint on the DNA), followed by a moderately long
> MD run in the NPT for Equilibration ( at 300 K). The .ncrst from
> minimization (only the coordinates) was to be used as the restart file for
> the heating step and the .ncrst from the heating (coordinates and
> velocities) was to be used as the restart for the Equilibration. Upon
> running the job as instructed and plotting the summary files on GNUplot, it
> was observed that initially the temperature rose as expected to 300 K
> within the 20 ps of heating simulation. However at the point of restart
> (i.e. right after the 20 ps mark) there is a sharp drop in Temperature by
> approx. 100 K and thereafter it again rose back to 300 K and stayed there
> the rest of the way.
> To further investigate this issue I ran the heating and Equilibration
> steps several times over with different break and restarts points and made
> the following observations :
> (i) The ensemble (NPT or NVT) did not matter and at each restart (
> irrespective of how many ps apart the restarts were done) there was a
> decrease in the temperature. I have tried breaking the 20 ps heating step
> into 2-3 steps and even in this I have observed the decrease at the restart
> points.
> (ii) The effect of this Temperature fall was observed in the other data
> like the Energy (Total, Potential and Kinetic - instead of reaching
> equilibrium values and fluctuating around the equilibrium values,these too
> decrease at the restart points),Pressure and Density.
> (iii) The .out file of the previous run shows final Temperature = 300 K
> and some energy values; however the next run's .out starts with different
> restart Temperature and energy values. This is observed at each break.
> From my Professor and Seniors I have come to know this issue was
> previously encountered on AMBER 18 but was not observed on running the same
> on AMBER 14.
> I understand that the Tutorial was made for the previous versions of AMBER
> and am confused regarding the cause of this error. Any guidance regarding
> this is much appreciated.
> I have herein attached the files for my last attempt at the simulation
> (this involves only the heating step; I have broken the heating step into 3
> parts : 1st is a 10 ps run with a 10 kcal/mol.A restraint on the DNA, 2nd a
> 6 ps run with a 5 kcal/mol.A restraint on DNA and the 3rd is a 4 ps run
> with 3 kcal/mol.A on the DNA.
> Thanking you,
> Diganta Dasgupta

AMBER mailing list

Received on Mon Sep 06 2021 - 07:00:02 PDT
Custom Search