Re: [modeller_usage] Do a loop refinment in order to mimic loop in a template
To: Miguel Lacerda <>, "" <>
Subject: Re: [modeller_usage] Do a loop refinment in order to mimic loop in a template
From: Modeller Caretaker <>
Date: Thu, 6 May 2021 16:44:49 -0700
Cc: Ana Oliveira <>
On 4/28/21 9:30 AM, Miguel Lacerda wrote:
We are modelling 80 isoforms of an enzyme superfamily. In the 1st step
we generated the models based on the templates that have higher sequence
identity. Now, we are trying to place a loop in a specific conformation.
This loop is very important for the enzyme catalysis. The catalytic
site is build up by a Calcium center coordinated with 6 residues. 4 of
those are located in the loop that does not have the desired
conformation. Our problem is due to the fact that the most of the
templates (selected based on their higher identity to the target) did
not have calcium in the X-ray structure which displaces the loop. What
we are trying now is to refine this problemic region in order to match
the only X-ray structure that was solved in presence of Calcium and
obtain the ideal conformation for the active enzyme. Do you have any
suggestions on how we could achieve this using modeller?
Modeller's loop refinement protocol is probably not going to help you
much here, since it will try to model the loop residues in an ab initio
fashion - it will not use any information from the templates, and it
does not consider specific interactions with ligands or ions. What you
probably want to do is build a regular comparative model using two
templates - one for the non-loop region (the majority of the sequence)
and another that just covers the loop region - and here you'd want to
use the structure that has the correct loop conformation. You'll
probably need to tweak this a little to make it work nicely, e.g. by
ensuring the two templates overlap a bit. See also
https://salilab.org/modeller/FAQ.html#1