Re: Atom velocity error in coarse grained simulation using MARTINI

From: Fotis Baltoumas (fbaltoumas_at_biol.uoa.gr)
Date: Sun Apr 01 2018 - 06:07:58 CDT

Hello Rakesh,

Have you performed enough minimization and equilibration? Often, that is
the source of many problems.  Since you're coarse-graining your system,
consider doing a few equilibration steps with gradually increasing time
step values and don't go straight to 10/20 fs from the start.  make sure
your system is heated properly, pressure is controlled well enough and
no geometry errors appear after all equilibration steps.

Having said the above, and assuming of course that you've minimized and
equilibrated your system properly, you may want to consider the
following points:

1.  How high is your time step value?  You may need to drop it to 10 fs
for namd CG simulations.

2.  How do you control your temperature and how high is it? You may need
to address that too.

3.  Are you using any extra restraints such as extrabonds or elastic
network restraints?  Those, when combined with a high time step may lead
to crashes.

4.  Are you trying to perform your microsecond-long simulation in one
run?  Try splitting it in smaller consecutive runs, each a few hundred
nanoseconds long, and work with restarts.

5. How many cpu cores are you using? You may need to experiment a bit to
optimize domain decomposition.  More cpus don't always lead to better
performance.

If you do a search in the mailing list you'll find that many people have
the same issue as you when it comes to NAMD and Martini coarse-grained
simulations.  Most of them stem from the manner through which the force
field (originally tailored for GROMACS) has been implemented.

Good luck,

Fotis Baltoumas

On 31/03/2018 18:16, Rakesh K wrote:
> Dear all,
>
> I have modelled a protein of 160 amino acids and simulated it for few
> 100ns in namd.
> Then I coarse grained the system using MARTINI and tried it to further
> simulate for some microseconds.
>
> The simulation was running and abruptly got crashed in between during
> the production run,
> something like this
>
>
> ERROR: Atom 420 velocity is 25965.1 12571.2 -6373.37 (limit is 1200,
> atom 35 of 70 on patch 159 pe 17)
> ERROR: Atoms moving too fast; simulation has become unstable (1 atoms
> on patch 159 pe 17).
> ERROR: Exiting prematurely; see error messages above.
> ERROR: Atom 496 velocity is -25706.5 -12714.7 6397.55 (limit is 1200,
> atom 31 of 67 on patch 158 pe 15)
> ERROR: Atom 476 velocity is -1050.43 -77.2906 -943.981 (limit is 1200,
> atom 63 of 67 on patch 158 pe 15)
> ERROR: Atoms moving too fast; simulation has become unstable (2 atoms
> on patch 158 pe 15).
>
> Can someone what is the error ??

-- 
*******************************************
Fotis A. Baltoumas
Bioinformatics Postgraduate Programme
Section of Cell Biology and Biophysics
Department of Biology, National & Kapodistrian University of Athens Panepistimiopolis, Athens 157 01, GREECE
    --------------------------------------
email : fbaltoumas_at_biol.uoa.gr
http://biophysics.biol.uoa.gr
http://bioinformatics.biol.uoa.gr
*******************************************
---
Αυτό το e-mail ελέγχθηκε για ιούς από το πρόγραμμα Avast antivirus.
https://www.avast.com/antivirus

This archive was generated by hypermail 2.1.6 : Tue Dec 31 2019 - 23:19:47 CST