My nucleic acid PDB file consists few non standard residues where the uracil, cytosine bases have been modified by adding methyl, phenol etc groups to them.
Can I use PRODRG to generate a topology for use in GROMACS?
I think better you build the topology of the building blocks of those residues by hand using the parameters of the GROMACS topologies for standard residues. This can be done easily and fast, and you will get better results.
PRODRG was developed mainly for drug design studies, in which you will not see delicate conformational processes such as folding. The same goes for double helixes and other nucleic acid secondary structure conformations.
Additionally, nucleic acids in GROMOS behave very badly in MD simulations, with the double helixes separating from each other in short simulations. So you would have a bad parameter for the start, and difficulties to compare to the natural residue. One more point: parameterization of residues in GROMOS is not simple