[gmx-users] problem with position restraints: X0 set to zero

[Julian Garrec]

Justin A. Lemkul wrote:
> Julian Garrec wrote:
>   
>> Dear GROMACS users,
>>
>> I am trying to equilibrate my system (monomeric protein in water) and I 
>> want to use position restraint on heavy atoms of the solute using the 
>> posre.itp file. For some reason, grompp applies correctly the force 
>> constant I want, but sets all the reference positions to zero. Of course 
>> I would like this ref. positions to be taken from my .gro file:
>>
>> ${GROMPP} -debug 10 -f inp.mdp -po out.mdp -c first.gro -r first.gro -p 
>> prm.top -o allprm.tpr -maxwarn 10
>>
>>
>> Here are the first line of my posre.itp file:
>>
>> [ position_restraints ]
>> ; atom  type      fx      fy      fz
>>     1     1 567 567 567
>>
>>
>> ... and the first line containing the "POSRES" flag in my tpr file 
>> (dumped with gmxdump):
>>
>> functype[441]=POSRES, pos0A=( 0.00000000e+00, 0.00000000e+00, 
>> 0.00000000e+00), fcA=( 5.67000000e+02, 5.67000000e+02, 5.67000000e+02), 
>> pos0B=( 0.00000000e+00, 0.00000000e+00, 0.00000000e+00), fcB=( 
>> 5.67000000e+02, 5.67000000e+02, 5.67000000e+02)
>>
>>
>> Using the -debug option with grompp didn't provide me further information.
>>
>> Does anybody know how to cure this problem ?
>>
>>     
>
> I don't know what pos0A is doing, but much earlier in the gmxdump output 
> should've been something along the lines of:
>
> posres_xA[    0]={ 3.19200e+00,  4.72500e+00,  2.66900e+00}
>
> This line holds the coordinates of each of the restrained atoms.
>
> -Justin
>   
Hello Justin,

Thanks for your answer. I wanted to perform further tests before 
answering you.

Actually, I do have such lines like:

posres_xA[    0]={ 3.19200e+00,  4.72500e+00,  2.66900e+00}

As far as I know, the lines containing the flag "functype" in the dumped 
.tpr file define all the functions that are used to compute the 
potential energy. Since position restraints involve additive terms of 
the form k(x-x_0)^2, I expected that both k and x0 are defined in the 
same "functype" line.

What I didn't say in my previous mail is that minimizing my system with 
such restraints still induces strong deviation from the experimental 
structure. Infact if I apply a huge force constant, I can see that the 
restraints are properly applied. It seems that their is a strong 
mismatching between the experimental structure and the FF I use...

Thanks again,

Julian

>   
>> Thanks,
>>
>> Julian
>>     
>
>