Re: problem at maintaining geometry coordinate

Date: Tue Aug 24 2010 - 05:51:07 CDT

Quoting Kwee Hong <>:

> ...
> I actually ran another simulation with the following adjustment in the
> parameter file:
> CAD SS 127.000 2.467
> CAD S1 127.000 2.56
> The force constant was taken from a literature which employ Charmm22, too.
> With this parameter, all the cadmium eventually clamped together at a single
> point even at the minimization stage. So, does this mean that I would need
> to adjust the force constant to get the desired coordination geometry? What
> would be the thumb of rule in adjusting the force constant so that I can
> eventually obtain the desired coordination geometry.
> There is another thing that puzzle me, out from the papers which I refer to
> in simulating the metallothionein, most of them just include the bond length
> and bond angle in their publication without stating the force constant they
> used in the simulation. Why does this happen since force constant affects so
> much in getting the right coordination geometry?
>> Besides, I don't see any entry in the "modpar" for bond angles or dihedrals
>> around the CAD atom. So, did you really set a formal bond between the
>> cadmium and the sulfur(s) in your topology file?!? [if you did, the program
>> should complain about the absence of the angle and dihedral terms...]
> Elsa, you have pointed out something which have puzzled me, too. I actually
> just patch the CAD-S bond. I believe this is not the formal bond right? So
> this would be the reason the program did not complain the angle and dihedral
> terms?

Assuming your TETR patch passed through psfgen without errors, then
yes, you set 4 formal bonds around each cadmium, but no angles nor
dihedrals. If you're using "exclude 1-4" (scaled or not) in your
simulations you'll be missing some (if not all) non-bonding
interactions between your Cd ions, and hence, with only the bond
parameter controlling your geometry around the ions, it's not
surprising that you get it wrong or 'clamped'.

Try re-defining your patch to include at least the angle terms (my
guess is that you'll probably get away without the dihedrals) ...


Elsa S. Henriques                     e-mail:
Investigadora Auxiliar
Biologia Estrutural e Computacional (@CNC)
e Grupo de Química Estrutural (@CCC)
Cv15 - RMB Lab
Departamento de Química - Universidade de Coimbra
3004-535 COIMBRA                         Portugal
Tel.: +351 23 9852080 - ext.417
Fax : +351 23 9827703

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:56:04 CST