Re: [modeller_usage] No template...

[Modeller Caretaker <modeller-care@ucsf.edu>]

On 3/1/11 6:31 AM, Marco Sette wrote:
> I have a protein consisting of a C-terminal and N-terminal region.
> While for the C-terminal one I have a good template either Modeller
> or Blast don't suggest templates for the N-terminal one. I have some
> suggestion using Modweb (which search performs Modweb?) but not very
> good. Could ou suggest me a website to search for a template? I
> looked at the mailing list (very long...) but seems this question has
> not been posed. Many thanks,

Since Modeller is a package for comparative modeling, if you don't have 
a template, you cannot model that region of your model.

I don't know what you mean by "Modeller ... doesn't suggest templates" 
since Modeller doesn't suggest templates for you. Perhaps you mean you 
are doing a sequence-sequence search to look for templates (step 1 of 
the basic tutorial) and you didn't find any? BLAST does essentially the 
same thing - except that it's using an approximation rather than "real" 
dynamic programming, so it runs faster - so is also unlikely to find a 
template if this search doesn't work.

ModWeb supplements this basic sequence-sequence search with more 
sensitive profile-based searches (it also tries PSI-BLAST, which is a 
similar method). But there are many other methods available to search 
for templates - e.g. HMM-based. One thing you could do is look at the 
best performers for the last CASP contest and see how they detect 
templates - some of those are available as web servers.

	Ben Webb, Modeller Caretaker
--
modeller-care@ucsf.edu             http://www.salilab.org/modeller/
Modeller mail list: http://salilab.org/mailman/listinfo/modeller_usage