With hundreds of tools for the protein structure prediction from primary sequence, can anyone provide a logical approach for higher accuracy?
While there are many methods for homology based modelling. If sequence does not have much sequence similarity say (
What are tools available to do knowledge based modelling?
I have CD data available. So I have some idea about Secondary Structure composition.
The problem with Rosetta web server/package is that you cannot recreate the performance they get in their own lab. The Rossetta performs best when the sampling space is very high for which they use the world wide grid. The server will not do that. However, if you have access to a large cluster you can download the Rossetta package and running it manually using a lot of parallel runs. The better alternative is to try the I-TASSER server/package which certainly is better than Rossetta for homology modeling and certainly comparable if not better for ab initio stuff. It is also much faster. You can access it at http://zhanglab.ccmb.med.umich.edu/I-TASSER/ . Here is the link to the performance evaluation of almost all structure prediction servers in the world during CASP9 http://zhanglab.ccmb.med.umich.edu/casp9/
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology