I have just used the I-TASSER server to predict the 3D structure of a fusion of 2 proteins. I specified as templates two pdb-s, corresponding to the structure of the first protein and to a close relative of the second protein, respectively. However, the output from I-TASSER provided me a model in which the structure of the second protein is very much different from the structure of the template provided and seems to be unrealistic. Could you please give me an advice on how to use I-TASSER (or another server) in order to get a reliable structural model for my fusion protein?