[gmx-users] Too many parameters in itp-file

[Mark Abraham]

On 8/11/2010 10:59 PM, Mikkel Vestergaard wrote:
> Hi,
> I have used an itp file which include following lines:
> [constraints]
> ;  i     j   funct   length
> ;   1     2    1       0.14
> ;   1     3    1       0.14
>
> ; for minimization purposes replace constraints by stiff bonds:
> ;
> ;[bonds]
> ;  i     j   funct   length   force const.
>    1     2    1       0.14    50000
>    1     3    1       0.14    50000

A better approach here is to use #ifdef FLEXIBLE, in the same manner 
that the water .itp files do. Now you can control your model with the 
"include = -DFLEXIBLE" line in your EM .mdp file, and not have to edit 
files (perhaps erroneously).

>
>
> It was intended to look like this (other lines commented out):
> [constraints]
> ;  i     j   funct   length
>    1     2    1       0.14
>    1     3    1       0.14
>
> ; for minimization purposes replace constraints by stiff bonds:
> ;
> ;[bonds]
> ;  i     j   funct   length   force const.
> ;   1     2    1       0.14    50000
> ;   1     3    1       0.14    50000
>
> So the only difference from the intended lines was the extra parameter 
> 50000. Does Gromacs use the extra parameter for something or is it 
> just ignored? (Got no warnings in simulations.)

You'd have to look at the code to be sure. GROMACS sometimes infers the 
form of the list of values by seeing how many of what type of value you 
have supplied. Here, it might guess that you might have been doing free 
energy perturbation between a state with length 0.14 and one with length 
50000. However, this would probably work for you, correctly and silently.

You can use gmxcheck to compare the contents of the correct and 
incorrect .tpr files to see what inference GROMACS took.

Mark