The template does not contain any structural information for those N-terminal residues, if you look at the pub entry 3Q8i (https://www.rcsb.org/structure/3Q8I), it starts with (M)TMKQL...

Are you sure the N-terminal residues of your sequence are actually part of your mature protein? They look like they could be a signal peptide.

If you look at the UniProt entry for the template (https://www.uniprot.org/uniprot/Q8T6R7), The mRNA for that protein does indeed code for a signal peptide very similar to your query sequence, MSVSVLVSSL VVLFCVQCLI EHIDGA, but this signal sequence would be cleaved off during protein production by the signal peptidase.

Dear Zeeshan

i'm agree with Annemarie Honegger that you have to consider the real protein fragment, and therefore, if present remove the predicted signal peptide from your sequence, because it will not be present in the mature protein.

However as a answer:

Swiss model (similarly to MODELLER) is an "homology modelling" based tool and he can build model only if a similar template is available and in case that the avaialble template cover only a portion of your protein you can built a model only for this part.

There are other software that are ab-initio (as Rosetta) or that are based on fragmented modelling (i-tasser)

in the video

https://www.blogger.com/video.g?token=AD6v5dyPGWsbPCAGPM4bhC7qrFmUL9TIfeArxzSqZ-ao_K6iC3z597YtqkSB2QIvBIxy9TVU52O9By7lhSuO-TFmnmd1-xRgRxnhW6uEon8lV-Lk-zMV2mZq1MT0o4gwaT38-wIRIIY

uploaded on my blog ProteoCool

you can find some basic information about those.

ciao

Manuele

If you still want to model that N-terminal (signal seq) residues you can give a try to ab-initio modelling using Robetta server (https://robetta.bakerlab.org/) or you can do a distance/contact-based protein structure prediction (http://raptorx.uchicago.edu/)