If you don't specify any bonded interactions, your molecule will simply fly apart.  Bonds and angles can be kept rigid with constraints (though angle constraints are somewhat unstable) and dihedrals can be restrained.  You need to specify all of these interactions if you want the internal degrees of freedom to be fixed, though strictly speaking dihedrals can only be restrained in Gromacs, not constrained at a fixed value.

Dear Justin Lemkul,

Thanks for your answer. What I have understood is: 1st I need to define all the force field parameters and I should apply the constrains. Do I need to define those parameters in the .itp file and How to select those parameters. For example, in the  distance_restraints directory I need to specify index, low, up1, up2 and fac. Similarly there are other parameters for angle and dihedral. How to select those parameters?

I'm not suggesting you use distance restraints; you shouldn't.  You need normal [bonds], [angles], and [dihedrals] directives in the topology.  I don't know how you should obtain the parameters for these; if the model was derived with totally rigid internal degrees of freedom, that's not something Gromacs can do (strictly speaking).  If you adapt parameters from another force field (given the fact that most ideal geometries for bonds and angles are almost the same, and if your dihedrals will be rigidified their parameters are fairly irrelevant), then appropriate use of the constraints keyword in the .mdp file (to keep bonds and angles rigid) and then proper use of [dihedral_restraints] with a strong force constant should be reasonable.  Whether or not such a model is stable or sensible is another matter.

Dear Justin Lemkul,

Following your suggestions, I have constructed the force field. The parameters have been taken from this paper:http://scitation.aip.org/content/aip/journal/jcp/138/20/10.1063/1.4807013. Now if I put "constraints = all-angles " in the .mdp file (the manual says that it constrains all the bonds and angles) it gives a warning:  "Molecule type 'ACEM' has 29 constraints. For stability and efficiency there should not be more constraints than internal number of degrees of freedom: 21". The same warning is also obtained for molecule type Coumarin 153.  Kindly suggest me how to deal with it. Also I am not putting the dihedral restraints  since I don't have parameters for it. Will it effect the dynamics? Or in general if I don't constrain the bonds, angles or constrain only the H bonds, then how will it effect the dynamics? Thanks in advance.                                 

When you get into the details, Gromacs isn't very efficient for this type of simulation.  You wind up with a massive number of constraints, which, as grompp tells you, you have more of than degrees of freedom.  Some software deals with this by dealing in internal coordinates, which can be more efficient.  Gromacs can't do this.

If you don't have dihedral restraints, then you'll get unphysical geometries as the other constraints are satisified.

Thank you for your suggestions Justin.

Yes I am currently using DL_POLY and constructing the force field with dl_field.

Dear Alavi,

          Thanks a lot. I am attaching my force field for Acetamide+NaSCN, constructed from Dl_FIELD. Kindly look at it and suggest me if any modification is needed.

Dear Saman,

                      Thanks a lot. Can you send me the modified force field file later.

Suman

Dear Saman

                  Thanks again. In the DL_FIELD, acetamide is already defined in DLPOLY_CHARMM.sf. I have given the same atom types for acetamide molecules and it has taken all the parameters from the charmm.sf and charmm.par files. Also you have put a question mark to a particular dihedral whose atomic indexes were not there in the original force file. I didn't understand it.

I find one fact interesting. Earlier when I used my force field file the density of the system was 1.28 gm/cc as compared to the experimental density 1.17 gm/cc but after modification the density goes higher to 1.34 gm/cc.

Dear Saman Alavi

@ Nagasree

For simulation of crystalline solids such as MOFs, you only need to provide the crystal structure and leave the bonded parameters (bond, angle and dihedral ) section empty. But you will need to restrain all atoms to their initial structure during the minimization step. As long as the initial structure is suitable for the given force field, the structure should not collapse during the dynamic even if you remove the harmonic restrains.

Javad

Javad Noroozi do you mean use position restraints and include crystal structure information in the pdb file?

For the attached files can you help me with input files for GROMACS?

In the answers above Justin Lemkul mentioned that " If you don't specify any bonded interactions, your molecule will simply fly apart. "

So I am not sure if I can leave bonded parameters empty.

I tried OBGMX and Avogadro to get UFF bonded parameters but they don't match and not sure which one is correct.

And Avogadro provide values in Charmm FF style, how can I change it to Amber FF or gromos FF style.

Also, how to specify dihedrals restrain?