AMBER Archive (2009)

Subject: [AMBER] RE: errors on remd

From: Ross Walker (ross_at_rosswalker.co.uk)
Date: Fri Jan 23 2009 - 10:38:06 CST


Dear Zgong,

>        I would like to do the remd, but some errors happened
>my mdin files like:
>Equilibration
> &cntrl
>   irest=1, ntx=5, 
>   nstlim=10000, dt=0.001,
>   ntt=3, gamma_ln=1.0,
>   temp0=XXXXX, ig=RANDOM_NUMBER,
>   ntc=2, ntf=2, nscm=1000,
>    ntb=0, igb=1,
>   cut=999.0, rgbmax=999.0,
>   ntpr=1000, ntwx=1000, ntwr=100000,
>   nmropt=1,
>   numexchg=1000,
> /
> &wt TYPE='END'
> /
>However, the error happened 
 
>input conversion error, unit 9, file /export/home/md/remd.rst.007
 
>the remd.rst.007 file like
                                                                            
   
>  389  0.1360000E+04  0.3756000E+03
>...
> -302.2943711 -56.2521559************  32.5331551 258.2713939-378.2040061
 
>before I do the remd, I have I have run 1ns simulations with each replica
to equilibrate them to their individual temperatures
>my amber version is amber9

Is this restart file from the 1ns equilibration or from some point during
the REMD? The *'s mean the value is too large to fit in the assigned space,
in this case more than 999.999999. This can occur because of a number of
reasons. Firstly if you start at a finite temperature your system can be
given a net impulse on its center of mass causing it to translate over time.
You should use nscm to remove the center of mass motion. The same thing can
occur if you have a thermostat. In your case this is a non-periodic system
so there is no wrapping involved. Your input shows that you have nscm set so
this shouldn't be occurring - at least your entire system shouldn't be
translating. However, it isn't clear if you set this for the equilibration
run so your system may already have translated a long way from the origin
during the equilibration. You should probably try to recenter your system
before doing the REMD - and or repeat the equilibration with nscm set.

The second option could be that your system has either dissociated or blown
up. This can happen if you are running implicit solvent with two entities
bound together, or with ions involved. This is effectively an infinitely
dilute system and so entropy will ultimately drive the species to
dissociate. This can happen very quickly if the binding is weak or they are
not in a good starting structure. I believe that the high temperatures in
REMD would also cause such systems to rapidly dissociate.

You should probably visualize all the trajectories to see what is happening
and why the coordinates are so large. Make sure your equilibration went well
before doing REMD.

Good luck,
Ross

/\
\/
|\oss Walker

| Assistant Research Professor |
| San Diego Supercomputer Center |
| Tel: +1 858 822 0854 | EMail:- ross_at_rosswalker.co.uk |
| http://www.rosswalker.co.uk | PGP Key available on request |

Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber