I want to predict the structure of my protein of interest, with 424 aminoacids of unknown structure. When i blast my protein against PDB database, only one match came with identity less than 40 %. When i did modelling with I-TASSER, the helix region is more than usual. My protein has 10 alpha helix structure, but model of I-TASSER shows more than 15 helix structure. My protein is a homodimer protein. Could anyone tell me how to proceed further and predict the structure of my protein?
Query coverage is 23% with max identity as 25%. I have also tried with increasing the expected threshold value, but still identity is less than 40%
23% of 424 is around 97 that is not a good choice to model with Homology based modeling. Even I-TASSER will also look for additional template to model whole of your protein. de novo or threading approach also doesnt seems justifiable. Try some other modeling server i.e. Swissmodel that might solve your problem.
I would recommend PHYRE2. It was very helpful for me several times.
I tried using Swiss model as you suggested, but it has stopped/failed, since there is no template found.
I doubt any of homology based modelling will solve your problem. I strongly suggest you Robetta server for structure modeling as it uses multiple approaches to predict protein structure. I think it will solve your problem as in the similar case I ever got success from Robetta server.
Dear,
As i suggested it is tough to get a proper template with such similarity short query coverage. Even if any tool will predict a structure, in my view the reliability score will be too low.
- Badges
- Science topic
- Similar topics
- Chemistry
- General Biochemistry
- Proteomics