Dear Tony and Antonello,
I too had the problem of not getting heteroatoms in my models when I had specified HETATM_IO = on. I sorted it out as follows...
1. Is the HETATM already parameterised in MODELER?
if so...
2. Have you specified the HETATM in the alignment file? I use MODELER through QUANTA, which puts heteroatoms into the PDB input files but doesn't automatically put heteroatoms into the sequence alignment, so I have to edit the file manually.
3. Have you aligned the HETATMs for template and model sequence? If the hetatm for template and model sequence are not equivalent then your hetatm occurs in the postion of what would be the next C-alpha atom in your chain.
It's very basic I know - but sometimes mistakes can happen.
I hope it works out for you both.
Derek.