Some protein files from PDB database have missing atoms, so when I use them with some molecular modeling programs, they complain about it. Is it there any web server that can fix/heal them? SOmebody told me about MODELLER, but as far as I know they do not have a server.
Use this interface for MODELLER, you should be able to upload your template (your incomplete protein structure) and paste the sequence alignment between your sequence and your protein (essentially, the same sequence repeated twice).
The only thing you need is the MODELLER key, which you can get in the MODELLER website upon registration for a license (for free).