I want to simulate the PNA molecule with Lammps. I made the input file for Lammps with the GaussView program and a homemade MATLAB script, but I do not have the force field coefficients (for bonds, angles, dihedrals).
I would suggest going to the web server CHARMMING.
It'll certainly help solving tricky force-field posers using all the available CHARMM-type force-fields.
Then, after doing some short simulation using CHARMMING, one could be ready to continue simulating the story as long as one needs on one's own computer(s).
Given that most of the parameters you need are already in most biomolecular force fields, save a few linkages, you should not use generic force fields like GAFF, CGenFF, or anything produced by PRODRG. There are probably only a few bonded parameters that need to be derived and verified, but most everything else can come from an existing force field, since you are talking about components of proteins and nucleic acids, which have been very well parametrized in several force fields.
You didn't mention what force field you intend to use.
You can use "moltemplate" package to build the structure and assign your force field to them to make the final data file for application in lammps. However, I'm not sure which force fields are supported by this package.
You might in effect use any temperature, but it would be reasonable, if you would like to investigate the temperature effects on your MD trajectory and its derivatives ...