[gmx-users] error comm-grps=Protein SOL

[Justin A. Lemkul]

M. Emal Alekozai wrote:
> Dear All,
> 
> I have in a mdrun.mdp file following settings: 
>    ; mode for center of mass motion removal =
>     comm-mode                = Linear
>    ; number of steps for center of mass motion removal =
>    nstcomm                  = 1
>    ; group(s) for center of mass motion removal =
>    comm-grps                = Protein SOL
> 
> This causes following warning/ error in the log file:
>    ...
>    Large VCM(group SOL): -0.00020, 0.00161, 0.00109, T-cm: inf
>    Large VCM(group rest): -0.00238, -0.01596, -0.10896, T-cm: inf
>    Large VCM(group SOL): 0.00006, 0.00002, -0.00005, T-cm: inf
>    Large VCM(group rest): -0.00009, 0.00483, 0.00106, T-cm: inf
>    ...
> 
> For a small testrun on 8 CPU (xtc: 2MB, trr: 22MB) I get 8 logfiles which are due to this warnings 160 Mb each.
> I looked the error Message up in the Gromacs mailinglist and found following question [2] but no solution :-( .
> The Gromacs manual also didn't helped, there was only one line on page 137.
> 
> 1: http://www.Gromacs.org/pipermail/gmx-users/2008-April/033316.html
> 
> If I use in the mdrun.mdp file one of the following options "comm-grps=Protein" or "comm-grps=SOL" or "comm-grps=SYSTEM" or ";comm-grps=" the warning does not appear in the log file.
> 
> 
> 
> I would like to know:
> 
> 1. What does the option "comm-grps = Protein SOL" physical mean?
> a) Does it mean:
> The protein and solvent atoms may have a initial drift. Merge all PROTEIN and SOL atoms into a single group. Calculate the center of mass of this merged group and reset the center of mass motion of this merged group in every step?
> 
> b) Or does it mean:
> For example because of initial setup "error" all the solvent atoms have a global drift in one direction and all the protein atoms have a drift in a other direction.
> Calculate for both groups seperatly the center of mass and remove for each group the center of mass motion seperatly.
> 

I believe (b) is the answer, more in the sense that if each group independently 
has its COM re-scaled, it can lead to collisions between the groups.  For 
systems with only lateral motion (bilayers, solid surfaces) this may not cause a 
problem, but for a freely-spinning and moving protein in solution, setting these 
groups independently is asking for trouble.

> 2. If the option "comm-grps = Protein SOL" physical means (1b) what do I have to change in my simulation setup to avoid the warnings/ error "T-cm: inf" in the log files?
> 

comm-grps = System (as you noted, it does not give the error!)

> 3. Are the messages "... T-cm: inf" in the logfile a error or just a warning. Can I just ignore the messages or do I have to adjust my inputfile?
> 

Anything infinite (inf) is a bad sign.  In real life, nothing in a finite system 
is infinite :)  Usually infinite values (forces, energies, temperatures, etc.) 
indicate collisions, explosions, and generally just nasty behavior.

> 4. If I'm allowed to ignore the warning, how do I tell Gromacs not to write them in the log file. For a middle size simulation the logfile will use up my entire diskspace before the simulation has ended. 
> 

If it appears in the log file, it's there for a reason.  Messages generally 
aren't just printed for kicks to clog up your disk :)

-Justin

> 
> Thanks
> Emal
> 

-- 
========================================

Justin A. Lemkul
Graduate Research Assistant
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================