Hi,
I am new to protein homology modeling and docking. I have a protein with UniProt ID P01106. I am trying to build a model using SWISS-MODEL and I am struck with template selection. What are the best criteria for selecting a template?
What are the basic steps in homology modeling and what are the best tools available freely for this procedure?
The basic assumption behind homology modelling is that similar sequence = similar structure. So the best criterion for template selection is sequence identity/similarity. The template identification tools in SWISS-MODEL are easy to find, in the menu under Tools.
There are multiple tutorials and guides online describing the steps in homology modelling. If you read the SWISS-MODEL workspace documentation, you'll find a link to a detailed protocol, published in Nature Protocols.
There are also several freely-available tools; as Rm stated already, MODELLER is very highly-regarded.
Other points:
Although sequence similarity is a key criterion, it is possible to model structures with low identity to a template. Again, consult the MODELLER tutorial.
As Rm stated, there are already models available for P01106. If you examine the Uniprot entry for this protein and scroll down to 3D structure databases, you'll find partial X-ray and NMR structures, as well as links to homology model databases. The link to the Protein Model Portal provides another online tool to re-model this protein if desired.
It would seem that only the C-terminus of this protein is present in both experimental structures and models; this may be due to disorder in the N-terminus.
As for structure viewing tools, I prefer SwissPDB's interface as zany as it looks because it has good functionality for quickly viewing sequence + structure at the same time. PyMOL http://www.pymol.org/ definitely is the most flexible and easy to generate a pretty figure.
In the bad old days, I used to use SwissPDB with PovRAY to make figures. Now I just bung it in PyMOL.