Re: [modeller_usage] Modeling a larger protein with multiple templates
To: nornik <>
Subject: Re: [modeller_usage] Modeling a larger protein with multiple templates
From: Modeller Caretaker <>
Date: Mon, 08 Feb 2010 13:32:43 -0800
Cc: "" <>
On 2/8/10 4:31 AM, nornik wrote:
I am modeling a large protein and I am using five different
templates on different regions of the protein.
...
The problem is getting the different templates in the same frame of
reference, since the alignment is created manually and not by
Modeller.
It doesn't make any difference how you created your alignment - the
restraints for modeling are all internal distances in individual
templates, so templates do not need to be structurally aligned with each
other (I'm assuming that's what you mean by "same frame of reference").
A Modeller-created alignment is no different in this regard from one you
make yourself. But this does mean that you need some overlap between
your templates so that the distances in the overlap regions are restrained.
The model sequence has very high identity to the core template so
what i would like to do is first model with only the core template
and then remodel the three large "inserts" with the other four
templates and still maintain the sulfur bridges.
The alignment you posted seems to have reasonable overlap, so it seems
like you should be able to do a regular modeling run, selecting all 5
templates as your 'knowns', and everything will just work. But you could
certainly build a model, then use the resulting model as a template for
a further run, if you really wanted to - you may just lose some accuracy.