I have a protein of sequence of 1215 amino acids. As no 3D structure available, therefore I tried to build its homology model however there is a section of sequence which doesn't align to any available PDB coordinates (amino acids ranging from 835 to 1215). Is there anyway to perform a modelling for this non-aligned sequence.
Try Phyre2, Phyre2 use multiple templates from PDB and for the regions were not templates are found are modeled by ab initio. Use the intensive mode of Phyre2.
http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index
@ Abhishek Kumar....I used it but wont find any modelling in the area of my concern.....I m interested to do the homology modelling in region of 1010-1050 region of the concerning sequence attached with this comment.
Please find it!!!!! Please help me
How u r not getting results for this region, add full length sequence and then share model.pdb!!!
Hi, I have tried with this server (http://ps2v2.life.nctu.edu.tw/index.php), but the final model is based in the combination of two templates of 44.79 and 37.84% identity, repectively . However, I´m waiting a result from other server that could be better.
Hi, Abhishek Kumar, have you an idea for this problem,I think that the best solution for it is looking for programs that combine varios templates since no full crystal structure are reported to date for this protein.
Regards
Thanks everybody for providing your concern!!
I also tried with other software but i got disorder region in that area of interest therefore that software is not able to perform any homology modelling over that region
(Software name Phyre2).
Again the problem is same of drawing a homology modelling of that particular area.
There is no software that can reliably model a region this large (~840-1080) that has no homology to an existing template structure. Furthermore, since it has strong disorder prediction, that emphasizes the fact that this region is highly flexible; even if you did model it, you would need to make hundreds of loop models to see which ensemble of conformations make the most sense with whatever hypothesis you're trying to test. It's a low-complexity stretch of sequence (lots of Glu, Ser, Thr), so it's not going to adopt any single conformation (unless it makes interactions with a binding partner).
You could fold it de novo ~10K times or more with Rosetta (David Baker lab) and examine clusters of Top Scoring structure decoys with all of the caveats that Sauder suggested above. That just gives you an arc of possible conformations. The disorder prediction is probably the most important one - try Disopred2 server (in addition to secondary structure predictions like PsiPred, SAM, and Jufo (Jens Meiler lab), etc.).
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology