Hi, I am trying to predict the structure of a 10 residue duplication in a large protein (450+ amino acids). The duplication results in the following peptide sequence change (where N is the duplicated region and X is unchanged sequence): XXXXXXX[NNNNNNNNNN]XXXXX -> XXXXXXX[NNNNNNNNNN][NNNNNNNNNN]XXXXX. There are no template structures for the duplication, but there are many template structures for the wild-type 10 residue sequence.
I have tried to model the duplication with I-TASSER but the predicted models are all essentially the same as the wild-type structure, but with a bunched up string in the region of the duplication (I assume this means that the modeling failed for the duplication and the algorithm relied heavily on homologous templates). I also tried to only model the 110aa domain of the protein containing the duplication, to make the modeling simpler, but this does not yelp.
Can anyone recommend a way to model the structure of this duplication? I assume homology modeling won't work here, can I specify in I-TASSER or elsewhere to use ab initio folding? Or can I use the structure from the wild-type 10aa region to inform the structure of the inserted duplication?
Any advice is very appreciated!
Thanks,
Alex