[gmx-users] Constraints specified in .mdp as well as in .top files

[Mark Abraham]

Vishwanath Dalvi wrote:
> Hi!
> 
> I have a question about a possible redundancy/conflict in specifying 
> bond-constraints.
> 
> I have in my simulation (among other things) a number of bonds - some of 
> which are harmonic while some are rigid (or constrained).
> 
> I specify the bonds in the .itp (or ultimately in the .top) files as 
> follows:
> 
> [ bonds ]
> ;  index1  index2    func             deq(nm)       k(kJ/mol/nm2)
>        1       2       1        0.1529000000   224262.4000000000
>        1      32       1        0.1810000000   185769.6000000000
>        2       3       1        0.1529000000   224262.4000000000
>        3       4       1        0.1529000000   224262.4000000000
> [ constraints ]
> ;  index1  index2   funct          length(nm)
>        1      30       1        0.1090000000
>        1      31       1        0.1090000000
>        2      28       1        0.1090000000
>        2      29       1        0.1090000000
> 
> However, it turns out that the .mdp file also requires one to specify 
> constraints in the form of:
> constraints      =  all-bonds/h-bonds etc
> 
> Do we have to specify constraints in the .mdp file if we already have 
> specified them in the .itp file?

I believe so. This is also true of restraints (which are a different 
concept in GROMACS).

Mark