I had a protein which contains more than 2000 amino acid in length, unfortunately the blastp results didn't show any identical templates and even the ab-intio server (robetta and itasser) cut off the max. residue length to 1500 amino acid. Is their any alternative method to build a complete theoretical model? If anyone has an idea, it would be great.
Attempting to model a 2000aa protein ab initio is not a good idea. You could try to identify individual domains and try to model those
Attempting to model a 2000aa protein ab initio is not a good idea. You could try to identify individual domains and try to model those
Hi,
As Andre said try doing pfam search.. Select the domain of your interest and find the aminoacid range of that domain. select those sequence and try to model it.
Along with I-tasser and rosetta you can also try phyre2.
Hai,
As Andre and Prakash said you can go for the Pfam search. and try to find out various domains in that protein, then you can model each of the domain by various ab initio techniques and after getting each domain try to thread all the domains according to the sequential order. you can get a model for the protein.
as already suggested by many, a domain search would be of great help. Try breaking down the sequence into random fragments (overlapping fragments will be good). search each fragment for similarity via BLAST or directly in pfam. that might help you get an idea of the domains. a bit of hit and trial should help to reach some conclusion.
@ thanks every one.. Its really great information form u alll
If you don't have any template, you cannot use Modeller for homology modelling.
But if you have a lot of sequences from different organisms, you can try to use EVfold, to predict correlation mutations and have a rough idea of your structure:
http://evfold.org/
Protein 3D Structure Computed from Evolutionary Sequence Variation:
http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0028766
Three-Dimensional Structures of Membrane Proteins from Genomic Sequencing:
http://www.cell.com/abstract/S0092-8674%2812%2900509-0
As already told by many, modeling the domain of the protein will give meaningful results when templates are unavailable. The uniprot entry of the protein will give u the details of the important segments, provided its a well annotated protein. Then proceed with ab initio structure prediction followed by molecular dynamics simulation to validate the structure
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools