[modeller_usage] Ligand parameters

I have been testing MODELLER with a ras-GDP template for inclusion of such a 'standard' nucleotide ligand as GDP. My most recent files are attached (.ali, .top, .log, and template/model pdb files). The problem: warnings (W>) in log file indicate lack of some parameters for GDP in the par.lib file, even though GDP is included as a residue in the res_type.lib. MODELLER adds some defaults values for the undefined parameters, but clearly does not handle the GDP accurately since, in the pdb model calculated, the guanine ring is nonplanar. I am not that familiar with the format of the par.lib file. Is there a reasonable way to add parameters to get the GDP ligand modeled properly?

Thanks,

Tom