Page 1 of 1

Top 10 Templates for Domain Assembly

Posted: Tue Sep 13, 2022 10:26 am
by rhajjo
I am using I-TASSER to generate a 3D protein model. The query amino acid sequence has a crystal structure in the PDB. However, this crystal structure isn't among the top templates for domain assembly. All top 10 templates picked by I-TASSER for domain assembly have sequences that are less similar to the query sequence (than the PDB structure that belong to the query sequence). Why is that?

I appreciate your thoughts on this.

Re: Top 10 Templates for Domain Assembly

Posted: Tue Sep 13, 2022 6:16 pm
by jlspzw
Hi

Our template library is PDB70, which is not an entire PDB, it takes a no-redundant PDB library with structure similarity cutoff TM-score=0.7. So it is possible that your native structure is not in the library. Since our library is not based on sequence similarity, so it is possible that the picked-up template has low sequence similarity with your native, but it should have structural similarity with your native, you can use TM-align check with it. Also if your sequence is a multi-domain protein, and each domain has a native structure, I-TASSER will try to pick up the longest templates to cover the largest domain, if no full-length templates. This is also why for multi-domain targets, some domain-level templates may not be shown in the top 10 templates. if you want to do domain assembly, and for each domain you know the structure, try our DEMO server (https://zhanggroup.org/DEMO/). if you want to do multi-domain protein modeling, try LOMETS3 (https://zhanggroup.org/LOMETS3) or I-TASSER-MTD (https://zhanggroup.org/I-TASSER-MTD/)