if you know the structure of your protein you can do it with the I-TASSER server.
You put your new sequence with the model of your protein and the server can create your structure. If you don't know the structure of your protein you can try the SWISSMODEL webserver. In this case its a homology modelling based on known structure from the PDB.
You may use 'align and mutate' option in Coot. Take the PDB of full length protein structure and use sequence of mutated (deletion) version of the protein for alignment. This will be hopefully, quickest way to do this.