Remove pairwise interactions between bonded atoms for RDF using lj/cut/coul/msm pair_style

I did the MD simulation using lj/cut/coul/msm pair_style to calculate RDF. It is reflected from the LAMMPS website that the pairwise interactions between atoms in the same bond, angle, or dihedral can not be removed for long-range coulomb interactions like coul/long, coul/msm. I have to remove the contribution of pairwise interaction between bonded atoms from my RDF (calculating in VMD after getting the dump file from LAMMPS). Please let me know how to do this for the lj/cut/coul/msm pair_style.
Thanks in advance.

You are confusing bond interactions during the simulation in LAMMPS with the contribution of bonded neighbours to the RDF during post-processing in VMD. These are two completely different phenomena.

You will have to see if VMD can remove the bonded neighbour contribution in its RDF calculation. Or you can simply open the RDF output text file and overwrite the bond-distance peaks with zero manually.

Thank you for the reply. As per my knowledge, there is no option to remove bonded atoms contribution in VMD.
I might be confused but what is the meaning of this description (copied from compute rdf command — LAMMPS documentation)----
“If you have a bonded system, then the settings of special_bonds command can remove pairwise interactions between atoms in the same bond, angle, or dihedral. This is the default setting for the special_bonds command, and means those pairwise interactions do not appear in the neighbor list. Because this fix uses a neighbor list, it also means those pairs will not be included in the RDF. This does not apply when using long-range coulomb interactions (coul/long, coul/msm, coul/wolf or similar.)”

I am using long-range coulomb interactions and getting contributions from the bonded atoms. Let’s say methanol, The RDF of C-O has a peak at CO bond length because of the intra-molecular contribution rather than the contribution of neabouring atoms which are not bonded.

The g(r) is an atomic property. It doesn’t care if atoms are bonded or not. If you have a C-O bond in your system, it should show up in the g(r)_CO.

The comment in the LAMMPS documentation is a warning for people that expect the bonded interaction peak and may not see it because LAMMPS uses the neighbor list for the g(r) calculation and such pairs may be removed from the neighbor list (but they must not for long-range coulomb and a few other conditions). VMD doesn’t use a neighbor list, but does a brute force O(N**2) loop over all pairs and thus computed the g(r) until a much larger value of r.