Dear Ibrahim M. Ibrahim I would say try I-Tasser https://zhanggroup.org/I-TASSER/ here you can insert the complete sequence in FASTA format (so including the missing 7 AA).

Only disadvantages are:

-You need to register (for free) so not a real problem

-It takes (depending on the length) nowadays days (or even more than a week) before you receive the result

Best regards

Hello Ibrahim:

1.) If your protein is a multimer and the loop is resolved into another subunit, you can take those coordinates for your region of interest after properly aligning each subunit.

2.) You can use Modeller, as long as you find a protein with a relatively high level of similarity to your protein of interest (https://salilab.org/modeller/).

3.) Similar to Modeller, you can use Swiss-Model (https://swissmodel.expasy.org/)

4.) Finally, you can use Alphafold directly, and compare how reasonable are the geometries of the loops it generates. Alphafold usually takes a long time to generate results, so you can use the database that has already been predicted for the PDB structures (https://alphafold.ebi.ac.uk/).

Best of luck!

ModLoop (ucsf.edu)

DaReUS-Loop: a web server to model loops in homology models (univ-paris-diderot.fr)

https://modbase.compbio.ucsf.edu/modloop/

Anna Stary-Weinzinger Zhuangwei Zhang Edward Francisco Méndez Otálvaro Rob Keller Thank you all for you help and suggestions

I uploaded this tutorial a week ago on Modeling missing loops of protein structure using Modeller interface in UCSF ChimeraX... hope it helps

https://www.youtube.com/watch?v=GV6jHhxR5T4&t=20s&ab_channel=Mohamedshehata